https://korshunov.ai/ko/article/12361-on-jeongcaeg-jeungryuyi-seonggongeun-gyosa-gyumoga-anin-sinho-pumjile-yeongu/korshunov.aiko2026-07-16T09:52:27.019621+00:00온-정책 증류의 성공은 교사 규모가 아닌 신호 품질에 의존한다는 연구 결과 https://korshunov.ai/es/article/12361-el-estudio-encuentra-que-el-exito-de-la-destilacion-on-policy-depende-de-la-de/korshunov.aies2026-07-16T09:52:27.019621+00:00El estudio encuentra que el éxito de la destilación on-policy depende de la calidad de la señal, no de la escala del maestro https://korshunov.ai/ar/article/12361-drs-tjd-n-njh-ltqtyr-l-lsys-y-tmd-l-jwd-lshr-wlys-hjm-lm-lm/korshunov.aiar2026-07-16T09:52:27.019621+00:00دراسة تجد أن نجاح التقطير على السياسة يعتمد على جودة الإشارة وليس حجم المعلم https://korshunov.ai/ja/article/12361-onhorishi-zheng-liu-nocheng-gong-hajiao-shi-moterunogui-mo-tehanakuxin-hao-pin/korshunov.aija2026-07-16T09:52:27.019621+00:00オンポリシー蒸留の成功は教師モデルの規模ではなく信号品質に依存することを研究が明らかに https://korshunov.ai/hi/article/12361-adhyyn-paataa-hai-ki-onn-ponlisii-ddisttileshn-kii-sphltaa-snket-kii-gunnvttaa/korshunov.aihi2026-07-16T09:52:27.019621+00:00अध्ययन पाता है कि ऑन-पॉलिसी डिस्टिलेशन की सफलता संकेत की गुणवत्ता पर निर्भर करती है, शिक्षक के स्केल पर नहीं https://korshunov.ai/ru/article/12361-issledovanie-pokazyvaet-chto-uspekh-on-polsinoi-distilliatsii-zavisit-ot-signala/korshunov.airu2026-07-16T09:52:27.019621+00:00Исследование показывает, что успех он-полсиной дистилляции зависит от качества сигнала, а не от масштаба учителя https://korshunov.ai/zh-hans/article/12361-yan-jiu-fa-xian-zai-xian-ce-lue-zheng-liu-de-cheng-gong-qu-jue-yu-xin-hao-zhi-er/korshunov.aizh-cn2026-07-16T09:52:27.019621+00:00研究发现，在线策略蒸馏的成功取决于信号质量，而非教师规模 https://korshunov.ai/id/article/12361-studi-menemukan-keberhasilan-distilasi-on-policy-bergantung-pada-kualitas-sinyal/korshunov.aiid2026-07-16T09:52:27.019621+00:00Studi menemukan keberhasilan distilasi on-policy bergantung pada kualitas sinyal, bukan skala guru https://korshunov.ai/en/article/12361-study-finds-on-policy-distillation-success-depends-on-signal-quality-not-teacher/korshunov.aien2026-07-16T09:52:27.019621+00:00Study finds on-policy distillation success depends on signal quality, not teacher scale https://korshunov.ai/fr/article/12361-une-etude-revele-que-le-succes-de-la-distillation-on-policy-depend-de-la-qualite/korshunov.aifr2026-07-16T09:52:27.019621+00:00Une étude révèle que le succès de la distillation on-policy dépend de la qualité du signal, et non de l'échelle du professeur https://korshunov.ai/pt-br/article/12361-estudo-descobre-que-o-sucesso-da-destilacao-on-policy-depende-da-qualidade-do-da/korshunov.aipt2026-07-16T09:52:27.019621+00:00Estudo descobre que o sucesso da destilação on-policy depende da qualidade do sinal, não da escala do professor https://korshunov.ai/zh-hans/article/12360-tong-guo-yin-pin-gan-zhi-da-xing-yu-yan-mo-xing-de-xi-li-du-fan-kui-gai-jin-wen/korshunov.aizh-cn2026-07-16T09:44:07.036661+00:00通过音频感知大型语言模型的细粒度反馈改进文本到音频的指令遵循 https://korshunov.ai/es/article/12360-mejora-del-seguimiento-de-instrucciones-en-la-generacion-de-texto-a-audio-fina/korshunov.aies2026-07-16T09:44:07.036661+00:00Mejora del seguimiento de instrucciones en la generación de texto a audio mediante retroalimentación fina de modelos de lenguaje grandes conscientes del audio https://korshunov.ai/ja/article/12360-o-teiodui-ying-da-gui-mo-yan-yu-moterukaranoxi-li-du-hui-teiohenozhi-shi-zhui/korshunov.aija2026-07-16T09:44:07.036661+00:00オーディオ対応大規模言語モデルからの細粒度フィードバックによるテキストからオーディオへの指示追従の改善 https://korshunov.ai/hi/article/12360-onddiyo-aveyr-bdde-bhaassaa-monddlon-se-suukssm-striiy-phiiddbaik-ke-maadhym-se/korshunov.aihi2026-07-16T09:44:07.036661+00:00ऑडियो-अवेयर बड़े भाषा मॉडलों से सूक्ष्म-स्तरीय फीडबैक के माध्यम से टेक्स्ट-टू-ऑडियो निर्देश पालन में सुधार https://korshunov.ai/en/article/12360-improving-text-to-audio-instruction-following-via-fine-grained-feedback-from/korshunov.aien2026-07-16T09:44:07.036661+00:00Improving Text-to-Audio Instruction Following via Fine-Grained Feedback from Audio-Aware Large Language Models https://korshunov.ai/ar/article/12360-thsyn-tb-lt-lymt-mn-lns-l-lswt-br-tgdhy-rj-dqyq-mn-nmdhj-lgwy-kbyr-w-y-llswt/korshunov.aiar2026-07-16T09:44:07.036661+00:00تحسين اتباع التعليمات من النص إلى الصوت عبر تغذية راجعة دقيقة من نماذج لغوية كبيرة واعية للصوت https://korshunov.ai/id/article/12360-meningkatkan-pengikut-instruksi-teks-ke-audio-melalui-umpan-balik-halus-dari/korshunov.aiid2026-07-16T09:44:07.036661+00:00Meningkatkan Pengikut Instruksi Teks-ke-Audio melalui Umpan Balik Halus dari Model Bahasa Besar Sadar Audio https://korshunov.ai/pt-br/article/12360-melhorando-o-seguimento-de-instrucoes-de-texto-para-audio-via-feedback-fino-de/korshunov.aipt2026-07-16T09:44:07.036661+00:00Melhorando o Seguimento de Instruções de Texto para Áudio via Feedback Fino de Modelos de Linguagem Grandes Conscientes de Áudio https://korshunov.ai/fr/article/12360-amelioration-du-respect-des-instructions-texte-vers-audio-grace-a-un-retour-fin/korshunov.aifr2026-07-16T09:44:07.036661+00:00Amélioration du respect des instructions texte-vers-audio grâce à un retour fin issu de grands modèles linguistiques sensibles au son https://korshunov.ai/ko/article/12360-odio-insig-daehyeong-eoneo-modelyi-sebunhwadoen-pideubaegeul-tonghan-tegseuteu/korshunov.aiko2026-07-16T09:44:07.036661+00:00오디오 인식 대형 언어 모델의 세분화된 피드백을 통한 텍스트-오디오 지시 따름 개선 https://korshunov.ai/ru/article/12360-uluchshenie-sledovaniia-instruktsiiam-v-zadachakh-preobrazovaniia-teksta-v-audio/korshunov.airu2026-07-16T09:44:07.036661+00:00Улучшение следования инструкциям в задачах преобразования текста в аудио с помощью детализированной обратной связи от аудио-ориентированных больших языковых моделей https://korshunov.ai/id/article/12359-attention-head-reweighting-mengadaptasi-llm-dengan-parameter-200-1000x-lebih/korshunov.aiid2026-07-16T09:37:16.013130+00:00Attention Head Reweighting mengadaptasi LLM dengan parameter 200-1000x lebih sedikit https://korshunov.ai/zh-hans/article/12359-zhu-yi-li-tou-zhong-jia-quan-shi-llmneng-gou-yi-200-1000bei-geng-shao-de-can-shu/korshunov.aizh-cn2026-07-16T09:37:16.013130+00:00注意力头重加权使LLM能够以200-1000倍更少的参数进行适配 https://korshunov.ai/hi/article/12359-dhyaan-hedd-riiveitting-llms-ko-200-1000x-km-pairaamiittr-ke-saath-anukuulit-hai/korshunov.aihi2026-07-16T09:37:16.013130+00:00ध्यान हेड रीवेइटिंग LLMs को 200-1000x कम पैरामीटर के साथ अनुकूलित करती है https://korshunov.ai/ja/article/12359-attention-head-reweightingha-harame-tashu-wo200-1000bei-xue-jian-shitellmwoshi/korshunov.aija2026-07-16T09:37:16.013130+00:00Attention Head Reweightingは、パラメータ数を200〜1000倍削減してLLMを適応させる https://korshunov.ai/ko/article/12359-attention-head-reweightingeun-paramiteo-sureul-200-1000bae-julyeo-llmeul/korshunov.aiko2026-07-16T09:37:16.013130+00:00Attention Head Reweighting은 파라미터 수를 200-1000배 줄여 LLM을 적응시킵니다 https://korshunov.ai/ar/article/12359-d-wzn-rwws-lntbh-ttkyf-m-lnmdhj-llgwy-lkbyr-bstkhdm-dd-ql-mn-lm-mlt-bmqdr-200-mr/korshunov.aiar2026-07-16T09:37:16.013130+00:00إعادة وزن رؤوس الانتباه تتكيف مع النماذج اللغوية الكبيرة باستخدام عدد أقل من المعاملات بمقدار 200-1000 مرة https://korshunov.ai/en/article/12359-attention-head-reweighting-adapts-llms-with-200-1000x-fewer-parameters/korshunov.aien2026-07-16T09:37:16.013130+00:00Attention Head Reweighting adapts LLMs with 200-1000x fewer parameters https://korshunov.ai/es/article/12359-el-reequilibrio-de-cabezales-de-atencion-adapta-llms-con-200-1000x-menos/korshunov.aies2026-07-16T09:37:16.013130+00:00El reequilibrio de cabezales de atención adapta LLMs con 200-1000x menos parámetros https://korshunov.ai/ru/article/12359-perevzveshivanie-atomov-vnimaniia-adaptiruet-llm-s-v-200-1000-raz-menshim/korshunov.airu2026-07-16T09:37:16.013130+00:00Перевзвешивание атомов внимания адаптирует LLM с в 200–1000 раз меньшим количеством параметров https://korshunov.ai/pt-br/article/12359-reponderacao-de-cabecas-de-atencao-adapta-llms-com-200-1000x-menos-parametros/korshunov.aipt2026-07-16T09:37:16.013130+00:00Reponderação de Cabeças de Atenção adapta LLMs com 200-1000x menos parâmetros https://korshunov.ai/fr/article/12359-attention-head-reweighting-adapte-les-llm-avec-200-a-1000-fois-moins-de/korshunov.aifr2026-07-16T09:37:16.013130+00:00Attention Head Reweighting adapte les LLM avec 200 à 1000 fois moins de paramètres https://korshunov.ai/ja/article/12358-yao-ji-shuo-ming-shu-sheng-cheng-notamenoqwenhe-suxiao-gui-mo-yan-yu-ninkuhou/korshunov.aija2026-07-16T09:27:52.517176+00:00薬剤説明書生成のためのQwenベース小規模言語モデルのトレーニング後アライメント https://korshunov.ai/ko/article/12358-yagmul-seolmyeongseo-saengseongeul-wihan-qwen-giban-sogyumo-eoneo-modelyi-sahu/korshunov.aiko2026-07-16T09:27:52.517176+00:00약물 설명서 생성을 위한 Qwen 기반 소규모 언어 모델의 사후 학습 정렬 https://korshunov.ai/es/article/12358-alineacion-post-entrenamiento-de-modelos-de-lenguaje-pequenos-basados-en-qwen-la/korshunov.aies2026-07-16T09:27:52.517176+00:00Alineación post-entrenamiento de modelos de lenguaje pequeños basados en Qwen para la generación de prospectos de medicamentos https://korshunov.ai/en/article/12358-post-training-alignment-of-qwen-based-small-language-models-for-medication/korshunov.aien2026-07-16T09:27:52.517176+00:00Post-training alignment of Qwen-based small language models for medication leaflet generation https://korshunov.ai/id/article/12358-penyelarasan-pasca-pelatihan-model-bahasa-kecil-berbasis-qwen-untuk-generasi/korshunov.aiid2026-07-16T09:27:52.517176+00:00Penyelarasan pasca-pelatihan model bahasa kecil berbasis Qwen untuk generasi brosur obat https://korshunov.ai/ar/article/12358-mhdh-m-b-d-ltdryb-llnmdhj-llgwy-lsgyr-lqym-l-qwen-ltwlyd-nshrt-ldwy/korshunov.aiar2026-07-16T09:27:52.517176+00:00محاذاة ما بعد التدريب للنماذج اللغوية الصغيرة القائمة على Qwen لتوليد نشرات الأدوية https://korshunov.ai/zh-hans/article/12358-ji-yu-qwende-xiao-xing-yu-yan-mo-xing-zai-yao-wu-shuo-ming-shu-sheng-cheng-zhong/korshunov.aizh-cn2026-07-16T09:27:52.517176+00:00基于Qwen的小型语言模型在药物说明书生成中的训练后对齐 https://korshunov.ai/fr/article/12358-alignement-post-entrainement-des-petits-modeles-de-langage-bases-sur-qwen-pour/korshunov.aifr2026-07-16T09:27:52.517176+00:00Alignement post-entraînement des petits modèles de langage basés sur Qwen pour la génération de notices médicales https://korshunov.ai/ru/article/12358-postobuchaiushchaia-adaptatsiia-malykh-iazykovykh-modelei-na-baze-qwen-dlia-po/korshunov.airu2026-07-16T09:27:52.517176+00:00Постобучающая адаптация малых языковых моделей на базе Qwen для генерации инструкций по применению лекарств https://korshunov.ai/hi/article/12358-dvaaon-ke-ptrk-jnreshn-ke-lie-qwen-aadhaarit-chotte-bhaassaa-monddl-kaa-postt/korshunov.aihi2026-07-16T09:27:52.517176+00:00दवाओं के पत्रक जनरेशन के लिए Qwen-आधारित छोटे भाषा मॉडल का पोस्ट-ट्रेनिंग अलाइनमेंट https://korshunov.ai/pt-br/article/12358-alinhamento-pos-treinamento-de-pequenos-modelos-de-linguagem-baseados-em-qwen-de/korshunov.aipt2026-07-16T09:27:52.517176+00:00Alinhamento pós-treinamento de pequenos modelos de linguagem baseados em Qwen para geração de bulas de medicamentos https://korshunov.ai/ru/article/12389-abliatsiia-olmo-core-pokazyvaet-chto-ispravlenie-indeksatsii-engram-suzhaet/korshunov.airu2026-07-16T09:21:03.972474+00:00Абляция OLMo-core показывает, что исправление индексации Engram сужает архитектурный разрыв https://korshunov.ai/en/article/12389-olmo-core-ablation-shows-engram-indexing-fixes-narrow-architecture-gap/korshunov.aien2026-07-16T09:21:03.972474+00:00OLMo-core ablation shows Engram indexing fixes narrow architecture gap https://korshunov.ai/ko/article/12389-olmo-core-abeulreisyeon-yeongu-engram-indegsingi-jobeun-akitegceo-gyeogcareul/korshunov.aiko2026-07-16T09:21:03.972474+00:00OLMo-core 아블레이션 연구: Engram 인덱싱이 좁은 아키텍처 격차를 수정함 https://korshunov.ai/zh-hans/article/12389-olmo-corexiao-rong-yan-jiu-biao-ming-engramsuo-yin-xiu-fu-suo-xiao-liao-jia-gou/korshunov.aizh-cn2026-07-16T09:21:03.972474+00:00OLMo-core消融研究表明，Engram索引修复缩小了架构差距 https://korshunov.ai/ar/article/12389-tzhr-drs-lzl-fy-olmo-core-n-fhrs-engram-tslh-lfjw-ldyq-fy-lbny/korshunov.aiar2026-07-16T09:21:03.972474+00:00تظهر دراسة الإزالة في OLMo-core أن فهرسة Engram تصلح الفجوة الضيقة في البنية https://korshunov.ai/es/article/12389-la-ablacion-de-olmo-core-muestra-que-las-correcciones-de-indexacion-de-engram-la/korshunov.aies2026-07-16T09:21:03.972474+00:00La ablación de OLMo-core muestra que las correcciones de indexación de Engram reducen la brecha arquitectónica https://korshunov.ai/pt-br/article/12389-ablacao-do-olmo-core-mostra-que-correcoes-de-indexacao-do-engram-estreitam-a/korshunov.aipt2026-07-16T09:21:03.972474+00:00Ablação do OLMo-core mostra que correções de indexação do Engram estreitam a lacuna arquitetônica https://korshunov.ai/ja/article/12389-olmo-corenoxiao-rong-shi-yan-te-engramintetsukusukaxia-ia-zheng/korshunov.aija2026-07-16T09:21:03.972474+00:00OLMo-coreの消融実験で、Engramインデックスが狭いアーキテクチャのギャップを修正 https://korshunov.ai/hi/article/12389-olmo-core-ebleshn-dikhaataa-hai-ki-engraam-inddeksing-phiks-aarkittekcrl-gaip-ko/korshunov.aihi2026-07-16T09:21:03.972474+00:00OLMo-core एब्लेशन दिखाता है कि एनग्राम इंडेक्सिंग फिक्स आर्किटेक्चरल गैप को संकर करता है https://korshunov.ai/id/article/12389-ablasi-olmo-core-menunjukkan-bahwa-pengindeksan-engram-memperbaiki-kesenjangan/korshunov.aiid2026-07-16T09:21:03.972474+00:00Ablasi OLMo-core menunjukkan bahwa pengindeksan Engram memperbaiki kesenjangan arsitektur yang sempit https://korshunov.ai/fr/article/12389-l-ablation-d-olmo-core-montre-que-l-indexation-engram-comble-un-ecart-etroit/korshunov.aifr2026-07-16T09:21:03.972474+00:00L'ablation d'OLMo-core montre que l'indexation Engram comble un écart architectural étroit https://korshunov.ai/ja/article/12388-firered-image-edit-v1-1noint8-convrotliang-zi-hua-woyu-sa-karikuesuto/korshunov.aija2026-07-16T09:12:32.003547+00:00FireRed Image Edit v1.1のINT8 ConvRot量子化をユーザーがリクエスト https://korshunov.ai/fr/article/12388-un-utilisateur-demande-la-quantification-convrot-int8-pour-firered-image-edit-v1/korshunov.aifr2026-07-16T09:12:32.003547+00:00Un utilisateur demande la quantification ConvRot INT8 pour FireRed Image Edit v1.1 https://korshunov.ai/en/article/12388-user-requests-int8-convrot-quantization-for-firered-image-edit-v1-1/korshunov.aien2026-07-16T09:12:32.003547+00:00User requests INT8 ConvRot quantization for FireRed Image Edit v1.1 https://korshunov.ai/pt-br/article/12388-usuario-solicita-quantizacao-convrot-int8-para-firered-image-edit-v1-1/korshunov.aipt2026-07-16T09:12:32.003547+00:00Usuário solicita quantização ConvRot INT8 para FireRed Image Edit v1.1 https://korshunov.ai/id/article/12388-pengguna-meminta-kuantisasi-convrot-int8-untuk-firered-image-edit-v1-1/korshunov.aiid2026-07-16T09:12:32.003547+00:00Pengguna meminta kuantisasi ConvRot INT8 untuk FireRed Image Edit v1.1 https://korshunov.ai/es/article/12388-usuario-solicita-cuantizacion-convrot-int8-para-firered-image-edit-v1-1/korshunov.aies2026-07-16T09:12:32.003547+00:00Usuario solicita cuantización ConvRot INT8 para FireRed Image Edit v1.1 https://korshunov.ai/hi/article/12388-upyogkrtaa-ne-firered-image-edit-v1-1-ke-lie-int8-convrot-kvaanttiikrnn-kaa/korshunov.aihi2026-07-16T09:12:32.003547+00:00उपयोगकर्ता ने FireRed Image Edit v1.1 के लिए INT8 ConvRot क्वांटीकरण का अनुरोध किया https://korshunov.ai/ar/article/12388-mstkhdm-ytlb-kmyw-convrot-int8-l-firered-image-edit-v1-1/korshunov.aiar2026-07-16T09:12:32.003547+00:00مستخدم يطلب كميّة ConvRot INT8 لـ FireRed Image Edit v1.1 https://korshunov.ai/zh-hans/article/12388-yong-hu-qing-qiu-wei-firered-image-edit-v1-1-ti-gong-int8-convrot-liang-hua/korshunov.aizh-cn2026-07-16T09:12:32.003547+00:00用户请求为 FireRed Image Edit v1.1 提供 INT8 ConvRot 量化 https://korshunov.ai/ru/article/12388-polzovatel-zaprashivaet-int8-kvantovanie-convrot-dlia-firered-image-edit-v1-1/korshunov.airu2026-07-16T09:12:32.003547+00:00Пользователь запрашивает INT8-квантование ConvRot для FireRed Image Edit v1.1 https://korshunov.ai/ko/article/12388-firered-image-edit-v1-1e-daehan-int8-convrot-yangjahwa-yoceong/korshunov.aiko2026-07-16T09:12:32.003547+00:00FireRed Image Edit v1.1에 대한 INT8 ConvRot 양자화 요청 https://korshunov.ai/ko/article/12387-gwahagjeog-ml-simyulreisyeonyi-deiteo-byeongmog-hyeonsang-ceorie-daehan-nonyi/korshunov.aiko2026-07-16T09:04:33.035316+00:00과학적 ML 시뮬레이션의 데이터 병목 현상 처리에 대한 논의 https://korshunov.ai/hi/article/12387-vaijnyaanik-ml-simuleshn-men-ddettaa-bonttlnek-ko-snbhaalne-pr-crcaa/korshunov.aihi2026-07-16T09:04:33.035316+00:00वैज्ञानिक ML सिमुलेशन में डेटा बॉटलनेक को संभालने पर चर्चा https://korshunov.ai/fr/article/12387-discussion-sur-la-gestion-des-goulots-d-etranglement-de-donnees-dans-les-ml/korshunov.aifr2026-07-16T09:04:33.035316+00:00Discussion sur la gestion des goulots d'étranglement de données dans les simulations ML scientifiques https://korshunov.ai/ja/article/12387-ke-xue-mlshimiyure-shiyonniokerute-tahotorunetsukunochu-li-niguan-suruyi-lun/korshunov.aija2026-07-16T09:04:33.035316+00:00科学MLシミュレーションにおけるデータボトルネックの処理に関する議論 https://korshunov.ai/zh-hans/article/12387-guan-yu-ke-xue-ji-qi-xue-xi-mo-ni-zhong-chu-li-shu-ju-ping-jing-de-tao-lun/korshunov.aizh-cn2026-07-16T09:04:33.035316+00:00关于科学机器学习模拟中处理数据瓶颈的讨论 https://korshunov.ai/es/article/12387-discusion-sobre-el-manejo-de-cuellos-de-botella-de-datos-en-simulaciones-de-ml/korshunov.aies2026-07-16T09:04:33.035316+00:00Discusión sobre el manejo de cuellos de botella de datos en simulaciones de ML científico https://korshunov.ai/ru/article/12387-obsuzhdenie-obrabotki-uzkikh-mest-v-dannykh-dlia-nauchnykh-ml-simuliatsii/korshunov.airu2026-07-16T09:04:33.035316+00:00Обсуждение обработки узких мест в данных для научных ML-симуляций https://korshunov.ai/en/article/12387-discussion-on-handling-data-bottlenecks-in-scientific-ml-simulation/korshunov.aien2026-07-16T09:04:33.035316+00:00Discussion on handling data bottlenecks in scientific ML simulation https://korshunov.ai/id/article/12387-diskusi-tentang-penanganan-bottleneck-data-dalam-simulasi-ml-ilmiah/korshunov.aiid2026-07-16T09:04:33.035316+00:00Diskusi tentang penanganan bottleneck data dalam simulasi ML ilmiah https://korshunov.ai/ar/article/12387-nqsh-hwl-m-lj-khtnqt-lbynt-fy-mhk-lt-lm-lly-l-lmy/korshunov.aiar2026-07-16T09:04:33.035316+00:00نقاش حول معالجة اختناقات البيانات في محاكاة التعلم الآلي العلمي https://korshunov.ai/pt-br/article/12387-discussao-sobre-o-tratamento-de-gargalos-de-dados-em-simulacoes-de-ml-cientifico/korshunov.aipt2026-07-16T09:04:33.035316+00:00Discussão sobre o tratamento de gargalos de dados em simulações de ML científico https://korshunov.ai/zh-hans/article/12386-llama-cpp-b10038-ban-ben-zai-shuo-ming-zhong-tian-jia-liao-guan-fang-wang-zhan/korshunov.aizh-cn2026-07-16T08:57:26.671890+00:00llama.cpp b10038 版本在说明中添加了官方网站链接 https://korshunov.ai/ru/article/12386-v-relize-llama-cpp-b10038-dobavlena-ssylka-na-ofitsialnyi-sait-v-primechaniiakh/korshunov.airu2026-07-16T08:57:26.671890+00:00В релизе llama.cpp b10038 добавлена ссылка на официальный сайт в примечаниях https://korshunov.ai/fr/article/12386-la-release-de-llama-cpp-b10038-ajoute-un-lien-vers-le-site-officiel-dans-les/korshunov.aifr2026-07-16T08:57:26.671890+00:00La release de llama.cpp b10038 ajoute un lien vers le site officiel dans les notes https://korshunov.ai/ja/article/12386-llama-cpp-b10038-riri-suno-tonigong-shi-uehusaitohenorinkuwozhui-jia/korshunov.aija2026-07-16T08:57:26.671890+00:00llama.cpp b10038 リリースノートに公式ウェブサイトへのリンクを追加 https://korshunov.ai/en/article/12386-llama-cpp-b10038-release-adds-official-website-link-to-notes/korshunov.aien2026-07-16T08:57:26.671890+00:00llama.cpp b10038 release adds official website link to notes https://korshunov.ai/id/article/12386-rilis-llama-cpp-b10038-menambahkan-tautan-situs-web-resmi-ke-catatan/korshunov.aiid2026-07-16T08:57:26.671890+00:00Rilis llama.cpp b10038 menambahkan tautan situs web resmi ke catatan https://korshunov.ai/pt-br/article/12386-lancamento-b10038-do-llama-cpp-adiciona-link-para-o-site-oficial-nas-notas/korshunov.aipt2026-07-16T08:57:26.671890+00:00Lançamento b10038 do llama.cpp adiciona link para o site oficial nas notas https://korshunov.ai/ar/article/12386-sdr-llama-cpp-b10038-ydyf-rbt-lmwq-lrsmy-l-lmlhzt/korshunov.aiar2026-07-16T08:57:26.671890+00:00إصدار llama.cpp b10038 يضيف رابط الموقع الرسمي إلى الملاحظات https://korshunov.ai/hi/article/12386-llama-cpp-b10038-riliij-men-notts-ko-aadhikaarik-vebsaaitt-link-joddaa-gyaa/korshunov.aihi2026-07-16T08:57:26.671890+00:00llama.cpp b10038 रिलीज़ में नोट्स को आधिकारिक वेबसाइट लिंक जोड़ा गया https://korshunov.ai/es/article/12386-la-version-b10038-de-llama-cpp-anade-enlace-al-sitio-web-oficial-en-las-notas/korshunov.aies2026-07-16T08:57:26.671890+00:00La versión b10038 de llama.cpp añade enlace al sitio web oficial en las notas https://korshunov.ai/ko/article/12386-llama-cpp-b10038-rilriseu-noteue-gongsig-websaiteu-ringkeu-cuga/korshunov.aiko2026-07-16T08:57:26.671890+00:00llama.cpp b10038 릴리스 노트에 공식 웹사이트 링크 추가 https://korshunov.ai/es/article/12357-modelos-de-difusion-discretos-un-marco-unificado-desde-la-tokenizacion-hasta-la/korshunov.aies2026-07-16T08:48:37.217185+00:00Modelos de difusión discretos: un marco unificado desde la tokenización hasta la generación https://korshunov.ai/fr/article/12357-modeles-de-diffusion-discrets-un-cadre-unifie-de-la-tokenisation-a-la-generation/korshunov.aifr2026-07-16T08:48:37.217185+00:00Modèles de diffusion discrets : un cadre unifié de la tokenisation à la génération https://korshunov.ai/ar/article/12357-nmdhj-lntshr-lmnfsl-tr-mwhd-mn-ltrmyz-l-ltwlyd/korshunov.aiar2026-07-16T08:48:37.217185+00:00نماذج الانتشار المنفصلة: إطار موحد من الترميز إلى التوليد https://korshunov.ai/ko/article/12357-isan-hwagsan-model-tokeunhwabuteo-saengseongggajiyi-tonghab-peureimweokeu/korshunov.aiko2026-07-16T08:48:37.217185+00:00이산 확산 모델: 토큰화부터 생성까지의 통합 프레임워크 https://korshunov.ai/ja/article/12357-li-san-kuo-san-moteru-to-kunhua-karasheng-cheng-matenotong-yi-hure-muwa-ku/korshunov.aija2026-07-16T08:48:37.217185+00:00離散拡散モデル：トークン化から生成までの統一フレームワーク https://korshunov.ai/zh-hans/article/12357-chi-san-kuo-san-mo-xing-cong-fen-ci-dao-sheng-cheng-de-tong-yi-kuang-jia/korshunov.aizh-cn2026-07-16T08:48:37.217185+00:00离散扩散模型：从分词到生成的统一框架 https://korshunov.ai/id/article/12357-model-difusi-diskrit-kerangka-terpadu-dari-tokenisasi-hingga-generasi/korshunov.aiid2026-07-16T08:48:37.217185+00:00Model Difusi Diskrit: Kerangka Terpadu dari Tokenisasi hingga Generasi https://korshunov.ai/hi/article/12357-ddiskriitt-ddiphyuujn-monddls-ttoknaaijeshn-se-jnreshn-tk-ek-ekiikrt-ddhaancaa/korshunov.aihi2026-07-16T08:48:37.217185+00:00डिस्क्रीट डिफ्यूजन मॉडल्स: टोकनाइजेशन से जनरेशन तक एक एकीकृत ढांचा https://korshunov.ai/pt-br/article/12357-modelos-de-difusao-discretos-um-framework-unificado-da-tokenizacao-a-geracao/korshunov.aipt2026-07-16T08:48:37.217185+00:00Modelos de Difusão Discretos: Um Framework Unificado da Tokenização à Geração https://korshunov.ai/en/article/12357-discrete-diffusion-models-a-unified-framework-from-tokenization-to-generation/korshunov.aien2026-07-16T08:48:37.217185+00:00Discrete Diffusion Models: A Unified Framework from Tokenization to Generation https://korshunov.ai/ru/article/12357-diskretnye-diffuzionnye-modeli-edinaia-kontseptualnaia-osnova-ot-tokenizatsii-do/korshunov.airu2026-07-16T08:48:37.217185+00:00Дискретные диффузионные модели: единая концептуальная основа от токенизации до генерации https://korshunov.ai/ko/article/12356-teugseong-giban-llm-paintyuningeun-keuroseu-rubeurig-esei-caejeom-seongneungeul/korshunov.aiko2026-07-16T08:42:32.943065+00:00특성 기반 LLM 파인튜닝은 크로스 루브릭 에세이 채점 성능을 향상시킵니다 https://korshunov.ai/pt-br/article/12356-ajuste-fino-de-llm-com-tracos-melhora-a-pontuacao-de-ensaios-em-multiplas/korshunov.aipt2026-07-16T08:42:32.943065+00:00Ajuste fino de LLM com traços melhora a pontuação de ensaios em múltiplas rúbricas https://korshunov.ai/ar/article/12356-dbt-lnmdhj-llgwy-lkbyr-llm-bstkhdm-lsmt-yhswn-tqyym-lmqlt-br-m-yyr-mt-dd/korshunov.aiar2026-07-16T08:42:32.943065+00:00ضبط النماذج اللغوية الكبيرة (LLM) باستخدام السمات يحسّن تقييم المقالات عبر معايير متعددة https://korshunov.ai/hi/article/12356-gunnon-ke-saath-llm-phaain-ttyuuning-krons-ruubrik-nibndh-skoring-ko-behtr-hai/korshunov.aihi2026-07-16T08:42:32.943065+00:00गुणों के साथ LLM फाइन-ट्यूनिंग क्रॉस-रूब्रिक निबंध स्कोरिंग को बेहतर बनाता है https://korshunov.ai/ja/article/12356-te-xing-woyong-itallmhuainchiyu-ninkuhakurosuru-huritsukuetsuseicai-dian-nogai/korshunov.aija2026-07-16T08:42:32.943065+00:00特性を用いたLLMファインチューニングはクロスルーブリックエッセイ採点の改善につながる https://korshunov.ai/es/article/12356-el-ajuste-fino-de-llm-con-rasgos-mejora-la-puntuacion-de-ensayos-segun-multiples/korshunov.aies2026-07-16T08:42:32.943065+00:00El ajuste fino de LLM con rasgos mejora la puntuación de ensayos según múltiples rúbricas https://korshunov.ai/ru/article/12356-tonkaia-nastroika-llm-s-chertami-uluchshaet-otsenku-esse-po-neskolkim-kriteriiam/korshunov.airu2026-07-16T08:42:32.943065+00:00Тонкая настройка LLM с чертами улучшает оценку эссе по нескольким критериям https://korshunov.ai/en/article/12356-llm-fine-tuning-with-traits-improves-cross-rubric-essay-scoring/korshunov.aien2026-07-16T08:42:32.943065+00:00LLM fine-tuning with traits improves cross-rubric essay scoring https://korshunov.ai/zh-hans/article/12356-shi-yong-te-zhi-wei-diao-da-yu-yan-mo-xing-ke-ti-sheng-kua-ping-fen-biao-zhun/korshunov.aizh-cn2026-07-16T08:42:32.943065+00:00使用特质微调大语言模型可提升跨评分标准作文评分 https://korshunov.ai/fr/article/12356-le-reglage-fin-des-llm-avec-des-traits-ameliore-la-notation-des-dissertations/korshunov.aifr2026-07-16T08:42:32.943065+00:00Le réglage fin des LLM avec des traits améliore la notation des dissertations selon plusieurs grilles https://korshunov.ai/id/article/12356-penyetelan-halus-llm-dengan-trait-meningkatkan-penilaian-esai-lintas-rubrik/korshunov.aiid2026-07-16T08:42:32.943065+00:00Penyetelan halus LLM dengan trait meningkatkan penilaian esai lintas rubrik https://korshunov.ai/ar/article/12355-ttb-gwrbny-lmbshr-m-yr-wnzm-mrj-y-lktb-trjmt-kyrtn-lsykh/korshunov.aiar2026-07-16T08:34:26.608272+00:00تتبع غورباني المباشر: معيار ونظام مرجعي لكتابة ترجمات كيرتان السيخ https://korshunov.ai/en/article/12355-live-gurbani-tracking-benchmark-and-reference-system-for-sikh-kirtan-captioning/korshunov.aien2026-07-16T08:34:26.608272+00:00Live Gurbani Tracking: Benchmark and Reference System for Sikh Kirtan Captioning https://korshunov.ai/hi/article/12355-laaiv-gurbaanii-ttraiking-sikh-kiirtn-kaipshning-ke-lie-bencmaark-aur-rephrens/korshunov.aihi2026-07-16T08:34:26.608272+00:00लाइव गुरबानी ट्रैकिंग: सिख कीर्तन कैप्शनिंग के लिए बेंचमार्क और रेफरेंस सिस्टम https://korshunov.ai/es/article/12355-seguimiento-en-vivo-de-gurbani-benchmark-y-sistema-de-referencia-para-la-de-sij/korshunov.aies2026-07-16T08:34:26.608272+00:00Seguimiento en vivo de Gurbani: Benchmark y sistema de referencia para la transcripción de Kirtan sij https://korshunov.ai/ko/article/12355-raibeu-gureubani-cujeog-sikeu-gyoyi-kireutan-jamag-jagseongeul-wihan-bencimakeu/korshunov.aiko2026-07-16T08:34:26.608272+00:00라이브 구르바니 추적: 시크 교의 키르탄 자막 작성을 위한 벤치마크 및 참조 시스템 https://korshunov.ai/zh-hans/article/12355-shi-shi-gu-er-ba-ni-zhui-zong-xi-ke-jiao-ji-er-tan-zi-mu-de-ji-zhun-yu-can-kao/korshunov.aizh-cn2026-07-16T08:34:26.608272+00:00实时古尔巴尼追踪：锡克教基尔坦字幕的基准与参考系统 https://korshunov.ai/id/article/12355-pelacakan-gurbani-langsung-benchmark-dan-sistem-referensi-untuk-subtitel-kirtan/korshunov.aiid2026-07-16T08:34:26.608272+00:00Pelacakan Gurbani Langsung: Benchmark dan Sistem Referensi untuk Subtitel Kirtan Sikh https://korshunov.ai/fr/article/12355-suivi-en-direct-de-la-gurbani-benchmark-et-systeme-de-reference-pour-la-sous-du/korshunov.aifr2026-07-16T08:34:26.608272+00:00Suivi en direct de la Gurbani : Benchmark et système de référence pour la sous-titration du Kirtan sikh https://korshunov.ai/ja/article/12355-sheng-kuruhanizhui-ji-shikujiao-tu-nokirutanzi-mu-zuo-cheng-notamenohenchima/korshunov.aija2026-07-16T08:34:26.608272+00:00生グルバニ追跡：シク教徒のキルタン字幕作成のためのベンチマークとリファレンスシステム https://korshunov.ai/ru/article/12355-otslezhivanie-gurbani-v-realnom-vremeni-etalonnaia-sistema-i-benchmark-dlia/korshunov.airu2026-07-16T08:34:26.608272+00:00Отслеживание Гурбани в реальном времени: эталонная система и бенчмарк для субтитрирования сикхского киртана https://korshunov.ai/pt-br/article/12355-acompanhamento-ao-vivo-de-gurbani-benchmark-e-sistema-de-referencia-para-de-sikh/korshunov.aipt2026-07-16T08:34:26.608272+00:00Acompanhamento ao vivo de Gurbani: Benchmark e sistema de referência para legendagem de Kirtan sikh https://korshunov.ai/pt-br/article/12354-devicesworld-apresenta-benchmark-em-larga-escala-para-colaboracao-de-agentes/korshunov.aipt2026-07-16T08:25:52.155648+00:00DevicesWorld apresenta benchmark em larga escala para colaboração de agentes entre dispositivos https://korshunov.ai/fr/article/12354-devicesworld-introduit-un-benchmark-a-grande-echelle-pour-la-collaboration-d/korshunov.aifr2026-07-16T08:25:52.155648+00:00DevicesWorld introduit un benchmark à grande échelle pour la collaboration d'agents multi-appareils https://korshunov.ai/ar/article/12354-devicesworld-tqdm-m-yran-ws-lntq-lt-wn-lwkl-br-ljhz/korshunov.aiar2026-07-16T08:25:52.155648+00:00DevicesWorld تقدم معيارًا واسع النطاق لتعاون الوكلاء عبر الأجهزة https://korshunov.ai/es/article/12354-devicesworld-presenta-un-benchmark-a-gran-escala-para-la-colaboracion-de-agentes/korshunov.aies2026-07-16T08:25:52.155648+00:00DevicesWorld presenta un benchmark a gran escala para la colaboración de agentes entre dispositivos https://korshunov.ai/hi/article/12354-devicesworld-ne-krons-ddivaais-ejentt-shyog-ke-lie-bdde-paimaane-pr-bencmaark/korshunov.aihi2026-07-16T08:25:52.155648+00:00DevicesWorld ने क्रॉस-डिवाइस एजेंट सहयोग के लिए बड़े पैमाने पर बेंचमार्क पेश किया https://korshunov.ai/en/article/12354-devicesworld-introduces-large-scale-benchmark-for-cross-device-agent/korshunov.aien2026-07-16T08:25:52.155648+00:00DevicesWorld introduces large-scale benchmark for cross-device agent collaboration https://korshunov.ai/ru/article/12354-devicesworld-predstavliaet-krupnomasshtabnyi-benchmark-dlia-kross-ustroistvennoi/korshunov.airu2026-07-16T08:25:52.155648+00:00DevicesWorld представляет крупномасштабный бенчмарк для кросс-устройственной агентной коллаборации https://korshunov.ai/zh-hans/article/12354-devicesworldtui-chu-kua-she-bei-zhi-neng-ti-xie-zuo-de-da-gui-mo-ji-zhun-ce-shi/korshunov.aizh-cn2026-07-16T08:25:52.155648+00:00DevicesWorld推出跨设备智能体协作的大规模基准测试 https://korshunov.ai/ko/article/12354-devicesworld-keuroseu-dibaiseu-eijeonteu-hyeobryeogeul-wihan-daegyumo-bencimakeu/korshunov.aiko2026-07-16T08:25:52.155648+00:00DevicesWorld, 크로스 디바이스 에이전트 협력을 위한 대규모 벤치마크 도입 https://korshunov.ai/id/article/12354-devicesworld-memperkenalkan-benchmark-skala-besar-untuk-kolaborasi-agen-lintas/korshunov.aiid2026-07-16T08:25:52.155648+00:00DevicesWorld memperkenalkan benchmark skala besar untuk kolaborasi agen lintas perangkat https://korshunov.ai/ja/article/12354-devicesworldkakurosutehaisue-shientoxie-li-notamenoda-gui-mo-henchima-kuwodao-ru/korshunov.aija2026-07-16T08:25:52.155648+00:00DevicesWorldがクロスデバイスエージェント協力のための大規模ベンチマークを導入 https://korshunov.ai/hi/article/12353-myag-snyojy-llm-ejentt-sisttm-ddijaain-krne-ke-lie-ek-graaph-aadhaarit-ddhaancaa/korshunov.aihi2026-07-16T08:19:17.484201+00:00MyAG: संयोज्य LLM एजेंट सिस्टम डिज़ाइन करने के लिए एक ग्राफ-आधारित ढांचा https://korshunov.ai/ko/article/12353-myag-johab-ganeunghan-llm-eijeonteu-siseutemeul-seolgyehagi-wihan-geuraepeu/korshunov.aiko2026-07-16T08:19:17.484201+00:00MyAG: 조합 가능한 LLM 에이전트 시스템을 설계하기 위한 그래프 기반 프레임워크 https://korshunov.ai/en/article/12353-myag-a-graph-based-framework-for-designing-composable-llm-agent-systems/korshunov.aien2026-07-16T08:19:17.484201+00:00MyAG: A graph-based framework for designing composable LLM agent systems https://korshunov.ai/ja/article/12353-myag-konho-sahurunallme-shientoshisutemuwoshe-ji-surutamenokurahuhe-sunohure-ku/korshunov.aija2026-07-16T08:19:17.484201+00:00MyAG: コンポーザブルなLLMエージェントシステムを設計するためのグラフベースのフレームワーク https://korshunov.ai/es/article/12353-myag-un-marco-basado-en-grafos-para-disenar-sistemas-de-agentes-llm-composables/korshunov.aies2026-07-16T08:19:17.484201+00:00MyAG: un marco basado en grafos para diseñar sistemas de agentes LLM composables https://korshunov.ai/ar/article/12353-myag-tr-ml-qym-l-lrswm-lbyny-ltsmym-nzm-wkl-llm-lqbl-lltrkyb/korshunov.aiar2026-07-16T08:19:17.484201+00:00MyAG: إطار عمل قائم على الرسوم البيانية لتصميم أنظمة وكلاء LLM القابلة للتركيب https://korshunov.ai/id/article/12353-myag-kerangka-berbasis-grafik-untuk-merancang-sistem-agen-llm-yang-dapat/korshunov.aiid2026-07-16T08:19:17.484201+00:00MyAG: Kerangka berbasis grafik untuk merancang sistem agen LLM yang dapat dikomposisikan https://korshunov.ai/pt-br/article/12353-myag-uma-estrutura-baseada-em-grafos-para-projetar-sistemas-de-agentes-llm/korshunov.aipt2026-07-16T08:19:17.484201+00:00MyAG: uma estrutura baseada em grafos para projetar sistemas de agentes LLM composáveis https://korshunov.ai/ru/article/12353-myag-grafovaia-platforma-dlia-proektirovaniia-kompozitsionnykh-sistem-agentov-na/korshunov.airu2026-07-16T08:19:17.484201+00:00MyAG: графовая платформа для проектирования композиционных систем агентов на основе LLM https://korshunov.ai/zh-hans/article/12353-myag-yi-chong-yong-yu-she-ji-ke-zu-he-llmzhi-neng-ti-xi-tong-de-ji-yu-tu-xing-de/korshunov.aizh-cn2026-07-16T08:19:17.484201+00:00MyAG：一种用于设计可组合LLM智能体系统的基于图形的框架 https://korshunov.ai/fr/article/12353-myag-un-cadre-base-sur-les-graphes-pour-concevoir-des-systemes-d-agents-llm/korshunov.aifr2026-07-16T08:19:17.484201+00:00MyAG : Un cadre basé sur les graphes pour concevoir des systèmes d'agents LLM composables https://korshunov.ai/ko/article/12352-eumseong-pyeonggareul-wihan-daegyumo-odio-eoneo-model-pansayi-peurotokol-sujun/korshunov.aiko2026-07-16T08:11:57.521819+00:00음성 평가를 위한 대규모 오디오 언어 모델 판사의 프로토콜 수준 단축 경로 감사 https://korshunov.ai/en/article/12352-auditing-protocol-level-shortcuts-in-large-audio-language-model-judges-for/korshunov.aien2026-07-16T08:11:57.521819+00:00Auditing Protocol-Level Shortcuts in Large Audio Language Model Judges for Speech Evaluation https://korshunov.ai/hi/article/12352-bolcaal-ke-muulyaankn-ke-lie-bdde-onddiyo-bhaassaa-monddl-jjon-men-prottokonl/korshunov.aihi2026-07-16T08:11:57.521819+00:00बोलचाल के मूल्यांकन के लिए बड़े ऑडियो-भाषा मॉडल जजों में प्रोटोकॉल-स्तरीय शॉर्टकट का ऑडिट https://korshunov.ai/pt-br/article/12352-auditoria-de-atalhos-em-nivel-de-protocolo-em-juizes-de-modelos-de-linguagem-na/korshunov.aipt2026-07-16T08:11:57.521819+00:00Auditoria de atalhos em nível de protocolo em juízes de modelos de linguagem grandes para áudio na avaliação de fala https://korshunov.ai/ja/article/12352-yin-sheng-ping-jia-niokeruda-gui-mo-o-teioyan-yu-moterupan-shi-tokatsutojian-cha/korshunov.aija2026-07-16T08:11:57.521819+00:00音声評価における大規模オーディオ言語モデル判事のプロトコルレベルのショートカット監査 https://korshunov.ai/fr/article/12352-audit-des-raccourcis-de-niveau-protocole-dans-les-juges-lalm-pour-l-evaluation/korshunov.aifr2026-07-16T08:11:57.521819+00:00Audit des raccourcis de niveau protocole dans les juges LALM pour l'évaluation de la parole https://korshunov.ai/ar/article/12352-mrj-lkhtsrt-l-mstw-lbrwtwkwl-fy-qd-nmdhj-llg-lswty-lkbyr-ltqyym-lklm/korshunov.aiar2026-07-16T08:11:57.521819+00:00مراجعة الاختصارات على مستوى البروتوكول في قضاة نماذج اللغة الصوتية الكبيرة لتقييم الكلام https://korshunov.ai/es/article/12352-auditoria-de-atajos-a-nivel-de-protocolo-en-jueces-de-modelos-de-lenguaje-para-y/korshunov.aies2026-07-16T08:11:57.521819+00:00Auditoría de atajos a nivel de protocolo en jueces de modelos de lenguaje grandes para audio y evaluación del habla https://korshunov.ai/zh-hans/article/12352-dui-yong-yu-yu-yin-ping-gu-de-da-xing-yin-pin-yu-yan-mo-xing-cai-pan-zhong-de-yi/korshunov.aizh-cn2026-07-16T08:11:57.521819+00:00对用于语音评估的大型音频语言模型裁判中的协议级捷径进行审计 https://korshunov.ai/ru/article/12352-audit-protokolnykh-uproshchenii-v-krupnykh-audio-iazykovykh-modeliakh-sudiakh/korshunov.airu2026-07-16T08:11:57.521819+00:00Аудит протокольных упрощений в крупных аудио-языковых моделях-судьях для оценки речи https://korshunov.ai/id/article/12352-mengaudit-pintasan-tingkat-protokol-dalam-hakim-model-bahasa-audio-besar-untuk/korshunov.aiid2026-07-16T08:11:57.521819+00:00Mengaudit Pintasan Tingkat Protokol dalam Hakim Model Bahasa Audio Besar untuk Evaluasi Ucapan https://korshunov.ai/pt-br/article/12351-sistema-bioasq-2026-usa-re-recuperacao-pragmatica-em-custo-e-fusao-de-selecao/korshunov.aipt2026-07-16T08:04:00.486755+00:00Sistema BioASQ 2026 usa re-recuperação pragmática em custo e fusão de seleção multi-modelo https://korshunov.ai/id/article/12351-sistem-bioasq-2026-menggunakan-re-retrieval-yang-pragmatis-secara-biaya-dan-fusi/korshunov.aiid2026-07-16T08:04:00.486755+00:00Sistem BioASQ 2026 menggunakan re-retrieval yang pragmatis secara biaya dan fusi-seleksi multi-model https://korshunov.ai/ar/article/12351-nzm-bioasq-2026-ystkhdm-d-lstrj-l-mly-mn-hyth-ltklf-wdmj-lkhtyr-mt-dd-lnmdhj/korshunov.aiar2026-07-16T08:04:00.486755+00:00نظام BioASQ 2026 يستخدم إعادة الاسترجاع العملي من حيث التكلفة ودمج الاختيار متعدد النماذج https://korshunov.ai/ko/article/12351-bioasq-2026-siseutemeun-biyong-hyoyuljeogin-jaegeomsaeg-mic-dajung-model-yunghab/korshunov.aiko2026-07-16T08:04:00.486755+00:00BioASQ 2026 시스템은 비용 효율적인 재검색 및 다중 모델 선택-융합 사용 https://korshunov.ai/zh-hans/article/12351-bioasq-2026xi-tong-cai-yong-cheng-ben-xiao-yi-xing-zhong-jian-suo-he-duo-mo-xing/korshunov.aizh-cn2026-07-16T08:04:00.486755+00:00BioASQ 2026系统采用成本效益型重检索和多模型选择-融合 https://korshunov.ai/en/article/12351-bioasq-2026-system-uses-cost-pragmatic-re-retrieval-and-multi-model-selection/korshunov.aien2026-07-16T08:04:00.486755+00:00BioASQ 2026 system uses cost-pragmatic re-retrieval and multi-model selection-fusion https://korshunov.ai/es/article/12351-el-sistema-bioasq-2026-utiliza-re-recuperacion-pragmatica-en-costos-y-fusion-de/korshunov.aies2026-07-16T08:04:00.486755+00:00El sistema BioASQ 2026 utiliza re-recuperación pragmática en costos y fusión de selección multi-modelo https://korshunov.ai/ru/article/12351-sistema-bioasq-2026-ispolzuet-pragmatichnoe-po-stoimosti-povtornoe-izvlechenie-i/korshunov.airu2026-07-16T08:04:00.486755+00:00Система BioASQ 2026 использует прагматичное по стоимости повторное извлечение и слияние выбора нескольких моделей https://korshunov.ai/hi/article/12351-bioasq-2026-sisttm-laagt-vyaavhaarik-punhpraapti-aur-bhu-monddl-cyn-snlyn-kaa/korshunov.aihi2026-07-16T08:04:00.486755+00:00BioASQ 2026 सिस्टम लागत-व्यावहारिक पुनःप्राप्ति और बहु-मॉडल चयन-संलयन का उपयोग करता है https://korshunov.ai/ja/article/12351-bioasq-2026-shisutemuhakosutoshi-yong-de-nazai-jian-suo-tomaruchimoteruxuan-ze/korshunov.aija2026-07-16T08:04:00.486755+00:00BioASQ 2026 システムはコスト実用的な再検索とマルチモデル選択融合を採用 https://korshunov.ai/fr/article/12351-le-systeme-bioasq-2026-utilise-une-re-recuperation-pragmatique-en-termes-de-cout/korshunov.aifr2026-07-16T08:04:00.486755+00:00Le système BioASQ 2026 utilise une ré-récupération pragmatique en termes de coût et une fusion par sélection multi-modèles https://korshunov.ai/es/article/12350-uts-evita-los-detectores-voight-kampff-de-eloquent-2026-mediante-cambios/korshunov.aies2026-07-16T07:54:11.278695+00:00UTS evita los detectores Voight-Kampff de ELOQUENT 2026 mediante cambios estructurales https://korshunov.ai/ru/article/12350-uts-obkhodit-detektory-voight-kampff-eloquent-2026-s-pomoshchiu-strukturnykh/korshunov.airu2026-07-16T07:54:11.278695+00:00UTS обходит детекторы Voight-Kampff ELOQUENT 2026 с помощью структурных сдвигов https://korshunov.ai/fr/article/12350-uts-contourne-les-detecteurs-voight-kampff-eloquent-2026-par-des-decalages/korshunov.aifr2026-07-16T07:54:11.278695+00:00UTS contourne les détecteurs Voight-Kampff ELOQUENT 2026 par des décalages structurels https://korshunov.ai/ja/article/12350-utskagou-zao-de-shihutoteeloquent-2026-voight-kampffjian-chu-qi-wohui-bi/korshunov.aija2026-07-16T07:54:11.278695+00:00UTSが構造的シフトでELOQUENT 2026 Voight-Kampff検出器を回避 https://korshunov.ai/zh-hans/article/12350-uts-tong-guo-jie-gou-pian-yi-rao-guo-eloquent-2026-voight-kampff-jian-ce-qi/korshunov.aizh-cn2026-07-16T07:54:11.278695+00:00UTS 通过结构偏移绕过 ELOQUENT 2026 Voight-Kampff 检测器 https://korshunov.ai/pt-br/article/12350-uts-contorna-detectores-voight-kampff-do-eloquent-2026-com-mudancas-estruturais/korshunov.aipt2026-07-16T07:54:11.278695+00:00UTS contorna detectores Voight-Kampff do ELOQUENT 2026 com mudanças estruturais https://korshunov.ai/ar/article/12350-uts-ytjwz-kshft-voight-kampff-lkhs-b-eloquent-2026-mn-khll-lthwlt-lhykly/korshunov.aiar2026-07-16T07:54:11.278695+00:00UTS يتجاوز كاشفات Voight-Kampff الخاصة بـ ELOQUENT 2026 من خلال التحولات الهيكلية https://korshunov.ai/en/article/12350-uts-bypasses-eloquent-2026-voight-kampff-detectors-with-structural-shifts/korshunov.aien2026-07-16T07:54:11.278695+00:00UTS bypasses ELOQUENT 2026 Voight-Kampff detectors with structural shifts https://korshunov.ai/id/article/12350-uts-melewati-detektor-voight-kampff-eloquent-2026-dengan-pergeseran-struktural/korshunov.aiid2026-07-16T07:54:11.278695+00:00UTS melewati detektor Voight-Kampff ELOQUENT 2026 dengan pergeseran struktural https://korshunov.ai/hi/article/12350-uts-snrcnaatmk-privrtnon-ke-saath-eloquent-2026-voight-kampff-ddittekttron-ko/korshunov.aihi2026-07-16T07:54:11.278695+00:00UTS संरचनात्मक परिवर्तनों के साथ ELOQUENT 2026 Voight-Kampff डिटेक्टरों को बायपास करता है https://korshunov.ai/ko/article/12350-uts-gujojeog-byeonhwaro-eloquent-2026-voight-kampff-tamjigireul-uhoe/korshunov.aiko2026-07-16T07:54:11.278695+00:00UTS, 구조적 변화로 ELOQUENT 2026 Voight-Kampff 탐지기를 우회 https://korshunov.ai/ru/article/12349-issledovanie-vyiavliaet-graduirovannye-pokazateli-znakomstva-sushchnostei-v/korshunov.airu2026-07-16T07:45:56.677798+00:00Исследование выявляет градуированные показатели знакомства сущностей в языковых моделях https://korshunov.ai/en/article/12349-study-finds-graded-entity-familiarity-readouts-in-language-models/korshunov.aien2026-07-16T07:45:56.677798+00:00Study finds graded entity-familiarity readouts in language models https://korshunov.ai/ar/article/12349-drs-tjd-qrt-tdryjy-lmlwfy-lkynt-fy-nmdhj-llg/korshunov.aiar2026-07-16T07:45:56.677798+00:00دراسة تجد قراءات تدريجية لمألوفية الكيانات في نماذج اللغة https://korshunov.ai/ja/article/12349-yan-yu-moteruniokeruduan-jie-de-naenteiteinoshou-xi-du-du-miqu-riwojian-chu-jiu/korshunov.aija2026-07-16T07:45:56.677798+00:00言語モデルにおける段階的なエンティティの熟悉度読み取りを検出する研究 https://korshunov.ai/fr/article/12349-une-etude-met-en-evidence-des-lectures-de-familiarite-graduees-des-entites-dans/korshunov.aifr2026-07-16T07:45:56.677798+00:00Une étude met en évidence des lectures de familiarité graduées des entités dans les modèles de langage https://korshunov.ai/hi/article/12349-adhyyn-men-bhaassaa-monddlon-men-krmik-ikaaii-pricittaa-riiddaautts-mile/korshunov.aihi2026-07-16T07:45:56.677798+00:00अध्ययन में भाषा मॉडलों में क्रमिक इकाई-परिचितता रीडआउट्स मिले https://korshunov.ai/ko/article/12349-eoneo-modeleseo-gyeceungjeog-entiti-cinsugdo-pandogeul-balgyeonhan-yeongu/korshunov.aiko2026-07-16T07:45:56.677798+00:00언어 모델에서 계층적 엔티티 친숙도 판독을 발견한 연구 https://korshunov.ai/es/article/12349-un-estudio-encuentra-lecturas-graduales-de-familiaridad-de-entidades-en-modelos/korshunov.aies2026-07-16T07:45:56.677798+00:00Un estudio encuentra lecturas graduales de familiaridad de entidades en modelos de lenguaje https://korshunov.ai/pt-br/article/12349-estudo-encontra-saidas-de-familiaridade-graduada-de-entidades-em-modelos-de/korshunov.aipt2026-07-16T07:45:56.677798+00:00Estudo encontra saídas de familiaridade graduada de entidades em modelos de linguagem https://korshunov.ai/zh-hans/article/12349-yan-jiu-fa-xian-yu-yan-mo-xing-zhong-cun-zai-fen-ji-shi-ti-shou-xi-du-du-shu/korshunov.aizh-cn2026-07-16T07:45:56.677798+00:00研究发现语言模型中存在分级实体熟悉度读数 https://korshunov.ai/id/article/12349-studi-menemukan-pembacaan-familiaritas-entitas-bertingkat-dalam-model-bahasa/korshunov.aiid2026-07-16T07:45:56.677798+00:00Studi menemukan pembacaan familiaritas entitas bertingkat dalam model bahasa https://korshunov.ai/zh-hans/article/12383-llama-cpp-b10037-yun-xu-tong-guo-pure-shou-dong-zhi-ding-zhang-liang-lei-xing/korshunov.aizh-cn2026-07-16T07:27:06.479658+00:00llama.cpp b10037 允许通过 --pure 手动指定张量类型 https://korshunov.ai/fr/article/12383-llama-cpp-b10037-permet-les-types-de-tenseurs-manuels-avec-pure/korshunov.aifr2026-07-16T07:27:06.479658+00:00llama.cpp b10037 permet les types de tenseurs manuels avec --pure https://korshunov.ai/ko/article/12383-llama-cpp-b10037eun-pure-peulraegeuro-sudong-tenseo-yuhyeongeul-heoyonghabnida/korshunov.aiko2026-07-16T07:27:06.479658+00:00llama.cpp b10037은 --pure 플래그로 수동 텐서 유형을 허용합니다 https://korshunov.ai/ar/article/12383-ysmh-llama-cpp-b10037-bnw-lmwtrt-lydwy-bstkhdm-pure/korshunov.aiar2026-07-16T07:27:06.479658+00:00يسمح llama.cpp b10037 بأنواع الموترات اليدوية باستخدام --pure https://korshunov.ai/en/article/12383-llama-cpp-b10037-allows-manual-tensor-types-with-pure/korshunov.aien2026-07-16T07:27:06.479658+00:00llama.cpp b10037 allows manual tensor types with --pure https://korshunov.ai/pt-br/article/12383-llama-cpp-b10037-permite-tipos-de-tensor-manuais-com-pure/korshunov.aipt2026-07-16T07:27:06.479658+00:00llama.cpp b10037 permite tipos de tensor manuais com --pure https://korshunov.ai/ja/article/12383-llama-cpp-b10037te-purehurakuniyorushou-dong-tensoruxing-nozhi-ding-kake-neng-ni/korshunov.aija2026-07-16T07:27:06.479658+00:00llama.cpp b10037で--pureフラグによる手動テンソル型の指定が可能に https://korshunov.ai/id/article/12383-llama-cpp-b10037-memungkinkan-tipe-tensor-manual-dengan-pure/korshunov.aiid2026-07-16T07:27:06.479658+00:00llama.cpp b10037 memungkinkan tipe tensor manual dengan --pure https://korshunov.ai/es/article/12383-llama-cpp-b10037-permite-tipos-de-tensor-manuales-con-pure/korshunov.aies2026-07-16T07:27:06.479658+00:00llama.cpp b10037 permite tipos de tensor manuales con --pure https://korshunov.ai/hi/article/12383-llama-cpp-b10037-pure-ke-saath-mainyual-ttensr-prkaaron-kii-anumti-detaa-hai/korshunov.aihi2026-07-16T07:27:06.479658+00:00llama.cpp b10037 --pure के साथ मैन्युअल टेंसर प्रकारों की अनुमति देता है https://korshunov.ai/ru/article/12383-llama-cpp-b10037-pozvoliaet-vruchnuiu-ukazyvat-tipy-tenzorov-s-pomoshchiu-pure/korshunov.airu2026-07-16T07:27:06.479658+00:00llama.cpp b10037 позволяет вручную указывать типы тензоров с помощью --pure https://korshunov.ai/hi/article/12348-memcon-llm-ejentts-ke-lie-anukuulit-memorii-prbndhn-siikhtaa-hai/korshunov.aihi2026-07-16T07:04:45.059335+00:00MemCon LLM एजेंट्स के लिए अनुकूलित मेमोरी प्रबंधन सीखता है https://korshunov.ai/pt-br/article/12348-memcon-aprende-gerenciamento-adaptativo-de-memoria-para-agentes-llm/korshunov.aipt2026-07-16T07:04:45.059335+00:00MemCon aprende gerenciamento adaptativo de memória para agentes LLM https://korshunov.ai/ru/article/12348-memcon-uchit-adaptivnomu-upravleniiu-pamiatiu-dlia-agentov-llm/korshunov.airu2026-07-16T07:04:45.059335+00:00MemCon учит адаптивному управлению памятью для агентов LLM https://korshunov.ai/zh-hans/article/12348-memcon-wei-llm-zhi-neng-ti-xue-xi-zi-gua-ying-nei-cun-guan-li/korshunov.aizh-cn2026-07-16T07:04:45.059335+00:00MemCon 为 LLM 智能体学习自适应内存管理 https://korshunov.ai/ko/article/12348-memconeun-llm-eijeonteureul-wihan-jeogeunghyeong-memori-gwanrireul/korshunov.aiko2026-07-16T07:04:45.059335+00:00MemCon은 LLM 에이전트를 위한 적응형 메모리 관리를 학습합니다 https://korshunov.ai/ar/article/12348-yt-lm-memcon-dr-ldhkr-ltkyfy-lwkl-llm/korshunov.aiar2026-07-16T07:04:45.059335+00:00يتعلم MemCon إدارة الذاكرة التكيفية لوكلاء LLM https://korshunov.ai/fr/article/12348-memcon-apprend-la-gestion-adaptative-de-la-memoire-pour-les-agents-llm/korshunov.aifr2026-07-16T07:04:45.059335+00:00MemCon apprend la gestion adaptative de la mémoire pour les agents LLM https://korshunov.ai/en/article/12348-memcon-learns-adaptive-memory-management-for-llm-agents/korshunov.aien2026-07-16T07:04:45.059335+00:00MemCon learns adaptive memory management for LLM agents https://korshunov.ai/ja/article/12348-memconhallme-shientonotamenoshi-ying-de-memoriguan-li-woxue-xi-suru/korshunov.aija2026-07-16T07:04:45.059335+00:00MemConはLLMエージェントのための適応的メモリ管理を学習する https://korshunov.ai/es/article/12348-memcon-aprende-gestion-adaptativa-de-memoria-para-agentes-llm/korshunov.aies2026-07-16T07:04:45.059335+00:00MemCon aprende gestión adaptativa de memoria para agentes LLM https://korshunov.ai/id/article/12348-memcon-belajar-manajemen-memori-adaptif-untuk-agen-llm/korshunov.aiid2026-07-16T07:04:45.059335+00:00MemCon belajar manajemen memori adaptif untuk agen LLM https://korshunov.ai/ru/article/12347-freimvork-cana-uluchshaet-sposobnost-agentov-llm-nakhodit-istoricheskie-analogii/korshunov.airu2026-07-16T06:56:41.235604+00:00Фреймворк CANA улучшает способность агентов LLM находить исторические аналогии https://korshunov.ai/hi/article/12347-cana-phremvrk-llm-ejentts-kii-aitihaasik-upmaaen-khojne-kii-kssmtaa-ko-behtr-hai/korshunov.aihi2026-07-16T06:56:41.235604+00:00CANA फ्रेमवर्क LLM एजेंट्स की ऐतिहासिक उपमाएँ खोजने की क्षमता को बेहतर बनाता है https://korshunov.ai/zh-hans/article/12347-canakuang-jia-ti-sheng-liao-da-yu-yan-mo-xing-zhi-neng-ti-xun-zhao-li-shi-lei-bi/korshunov.aizh-cn2026-07-16T06:56:41.235604+00:00CANA框架提升了大语言模型智能体寻找历史类比的能力 https://korshunov.ai/ja/article/12347-canahure-muwa-kukallme-shientonoli-shi-de-lei-tui-fa-jian-neng-li-woxiang-shang/korshunov.aija2026-07-16T06:56:41.235604+00:00CANAフレームワークがLLMエージェントの歴史的類推発見能力を向上 https://korshunov.ai/es/article/12347-el-marco-cana-mejora-la-capacidad-de-los-agentes-llm-para-encontrar-analogias/korshunov.aies2026-07-16T06:56:41.235604+00:00El marco CANA mejora la capacidad de los agentes LLM para encontrar analogías históricas https://korshunov.ai/id/article/12347-kerangka-kerja-cana-meningkatkan-kemampuan-agen-llm-dalam-menemukan-analogi/korshunov.aiid2026-07-16T06:56:41.235604+00:00Kerangka kerja CANA meningkatkan kemampuan agen LLM dalam menemukan analogi historis https://korshunov.ai/ar/article/12347-tr-ml-cana-yhsn-qdr-wkl-lnmdhj-llgwy-lkbyr-l-l-thwr-l-ltnzrt-ltrykhy/korshunov.aiar2026-07-16T06:56:41.235604+00:00إطار عمل CANA يحسن قدرة وكلاء النماذج اللغوية الكبيرة على العثور على التناظرات التاريخية https://korshunov.ai/pt-br/article/12347-framework-cana-melhora-a-capacidade-de-agentes-llm-de-encontrar-analogias/korshunov.aipt2026-07-16T06:56:41.235604+00:00Framework CANA melhora a capacidade de agentes LLM de encontrar analogias históricas https://korshunov.ai/ko/article/12347-cana-peureimweokeuga-llm-eijeonteuyi-yeogsajeog-yucu-balgyeon-neungryeog/korshunov.aiko2026-07-16T06:56:41.235604+00:00CANA 프레임워크가 LLM 에이전트의 역사적 유추 발견 능력 향상 https://korshunov.ai/en/article/12347-cana-framework-improves-llm-agents-ability-to-find-historical-analogies/korshunov.aien2026-07-16T06:56:41.235604+00:00CANA framework improves LLM agents' ability to find historical analogies https://korshunov.ai/fr/article/12347-le-cadre-cana-ameliore-la-capacite-des-agents-llm-a-trouver-des-analogies/korshunov.aifr2026-07-16T06:56:41.235604+00:00Le cadre CANA améliore la capacité des agents LLM à trouver des analogies historiques https://korshunov.ai/fr/article/12346-canon-transforme-le-consensus-en-supervision-dense-au-niveau-des-tokens-pour-l/korshunov.aifr2026-07-16T06:50:27.057216+00:00CANON transforme le consensus en supervision dense au niveau des tokens pour l'auto-distillation sans étiquette https://korshunov.ai/ko/article/12346-canoni-reibeul-eobsneun-jagi-jeungryu-label-free-self-distillation-reul-wihan/korshunov.aiko2026-07-16T06:50:27.057216+00:00CANON이 레이블 없는 자기 증류(label-free self-distillation)를 위한 밀집 토큰 수준 감독으로 합의를 전환하다 https://korshunov.ai/id/article/12346-canon-mengubah-konsensus-menjadi-pengawasan-tingkat-token-yang-padat-untuk-bebas/korshunov.aiid2026-07-16T06:50:27.057216+00:00CANON mengubah konsensus menjadi pengawasan tingkat token yang padat untuk distilasi mandiri bebas label https://korshunov.ai/es/article/12346-canon-convierte-el-consenso-en-supervision-densa-a-nivel-de-token-para-auto-sin/korshunov.aies2026-07-16T06:50:27.057216+00:00CANON convierte el consenso en supervisión densa a nivel de token para auto-distilación sin etiquetas https://korshunov.ai/pt-br/article/12346-canon-transforma-consenso-em-supervisao-densa-por-token-para-auto-distilacao-sem/korshunov.aipt2026-07-16T06:50:27.057216+00:00CANON transforma consenso em supervisão densa por token para auto-distilação sem rótulos https://korshunov.ai/ja/article/12346-canonkakonsensasuworaherunashizi-ji-zheng-liu-notamenomi-nato-kunreherunojiao/korshunov.aija2026-07-16T06:50:27.057216+00:00CANONがコンセンサスをラベルなし自己蒸留のための密なトークンレベルの教師信号に変える https://korshunov.ai/en/article/12346-canon-turns-consensus-into-dense-token-level-supervision-for-label-free-self/korshunov.aien2026-07-16T06:50:27.057216+00:00CANON turns consensus into dense token-level supervision for label-free self-distillation https://korshunov.ai/zh-hans/article/12346-canonjiang-gong-shi-zhuan-hua-wei-wu-biao-qian-zi-zheng-liu-zhong-de-mi-ji-jian/korshunov.aizh-cn2026-07-16T06:50:27.057216+00:00CANON将共识转化为无标签自蒸馏中的密集token级监督 https://korshunov.ai/hi/article/12346-canon-lebl-phrii-selph-ddisttileshn-ke-lie-knsenss-ko-ghntv-ttokn-str-kii-men/korshunov.aihi2026-07-16T06:50:27.057216+00:00CANON लेबल-फ्री सेल्फ-डिस्टिलेशन के लिए कंसेंसस को घनत्व टोकन-स्तर की निगरानी में बदलता है https://korshunov.ai/ar/article/12346-thwl-canon-ljm-l-shrf-kthyf-l-mstw-lrmwz-lltdryb-ldhty-lkhly-mn-ltsmyt/korshunov.aiar2026-07-16T06:50:27.057216+00:00تحول CANON الإجماع إلى إشراف كثيف على مستوى الرموز للتدريب الذاتي الخالي من التسميات https://korshunov.ai/ru/article/12346-canon-prevrashchaet-konsensus-v-plotnoe-tokeno-urovnevoe-obuchenie-dlia-bez/korshunov.airu2026-07-16T06:50:27.057216+00:00CANON превращает консенсус в плотное токено-уровневое обучение для самодистилляции без меток https://korshunov.ai/pt-br/article/12345-armacao-de-agente-autoevolutiva-melhora-desempenho-de-modelo-congelado-em-9-a-pp/korshunov.aipt2026-07-16T06:41:54.726650+00:00Armação de agente autoevolutiva melhora desempenho de modelo congelado em 9 a 15,5 pp https://korshunov.ai/ar/article/12345-d-wkyl-dhty-lttwr-thsn-d-lnmwdhj-lmjmd-bnsb-9-l-15-5-nqt-mywy/korshunov.aiar2026-07-16T06:41:54.726650+00:00أداة وكيل ذاتي التطور تحسن أداء النموذج المجمد بنسبة 9 إلى 15.5 نقطة مئوية https://korshunov.ai/ko/article/12345-jagijinhwa-eijeonteu-haneseuga-donggyeol-model-seongneungeul-9-15-5pp-hyangsang/korshunov.aiko2026-07-16T06:41:54.726650+00:00자기진화 에이전트 하네스가 동결 모델 성능을 9~15.5pp 향상 https://korshunov.ai/fr/article/12345-un-harnais-d-agent-auto-evolutif-ameliore-les-performances-des-modeles-figes-de/korshunov.aifr2026-07-16T06:41:54.726650+00:00Un harnais d'agent auto-évolutif améliore les performances des modèles figés de 9 à 15,5 pp https://korshunov.ai/id/article/12345-harness-agen-yang-berevolusi-sendiri-meningkatkan-kinerja-model-beku-sebesar-9/korshunov.aiid2026-07-16T06:41:54.726650+00:00Harness agen yang berevolusi sendiri meningkatkan kinerja model beku sebesar 9 hingga 15,5 pp https://korshunov.ai/zh-hans/article/12345-zi-jin-hua-zhi-neng-ti-kuang-jia-shi-dong-jie-mo-xing-xing-neng-ti-sheng-9zhi-15/korshunov.aizh-cn2026-07-16T06:41:54.726650+00:00自进化智能体框架使冻结模型性能提升9至15.5个百分点 https://korshunov.ai/ru/article/12345-samosovershenstvuiushchiisia-agentnyi-karkas-uluchshaet-pokazateli-zamorozhennoi/korshunov.airu2026-07-16T06:41:54.726650+00:00Самосовершенствующийся агентный каркас улучшает показатели замороженной модели на 9–15,5 процентных пункта https://korshunov.ai/en/article/12345-self-evolving-agent-harness-improves-frozen-model-performance-by-9-to-15-5pp/korshunov.aien2026-07-16T06:41:54.726650+00:00Self-evolving agent harness improves frozen model performance by 9 to 15.5pp https://korshunov.ai/ja/article/12345-zi-ji-jin-hua-xing-e-shientoha-nesukadong-jie-moterunoxing-neng-wo9-15-5ppxiang/korshunov.aija2026-07-16T06:41:54.726650+00:00自己進化型エージェントハーネスが凍結モデルの性能を9〜15.5pp向上 https://korshunov.ai/es/article/12345-el-arnes-de-agente-autoevolutivo-mejora-el-rendimiento-del-modelo-congelado-9-y/korshunov.aies2026-07-16T06:41:54.726650+00:00El arnés de agente autoevolutivo mejora el rendimiento del modelo congelado entre 9 y 15,5 pp https://korshunov.ai/hi/article/12345-sv-vikaasshiil-ejentt-haarnes-phrojn-monddl-kii-prdrshn-kssmtaa-ko-9-se-15-5pp/korshunov.aihi2026-07-16T06:41:54.726650+00:00स्व-विकासशील एजेंट हार्नेस फ्रोजन मॉडल की प्रदर्शन क्षमता को 9 से 15.5pp तक बढ़ाता है https://korshunov.ai/pt-br/article/12344-problema-do-oraculo-de-teste-em-corpora-sinteticos-de-llm-as-judge-distorcao-e/korshunov.aipt2026-07-16T06:34:26.265387+00:00Problema do Oráculo de Teste em Corpora Sintéticos de LLM-as-Judge: Desaparecimento, Distorção e um Protocolo de Validação https://korshunov.ai/hi/article/12344-sinthettik-llm-as-judge-konrpraa-men-ttestt-orekl-smsyaa-gaayb-honaa-vikrti-aur/korshunov.aihi2026-07-16T06:34:26.265387+00:00सिंथेटिक LLM-as-Judge कॉर्परा में टेस्ट ओरेकल समस्या: गायब होना, विकृति और एक वैलिडेशन प्रोटोकॉल https://korshunov.ai/ja/article/12344-he-cheng-llm-as-judgeko-hasuniokerutesutoorakuruwen-ti-xiao-shi-wai-qu-oyohijian/korshunov.aija2026-07-16T06:34:26.265387+00:00合成LLM-as-Judgeコーパスにおけるテストオラクル問題：消失、歪曲、および検証プロトコル https://korshunov.ai/ru/article/12344-problema-testovogo-orakula-v-sinteticheskikh-korpusakh-llm-as-judge-iskazhenie-i/korshunov.airu2026-07-16T06:34:26.265387+00:00Проблема тестового оракула в синтетических корпусах LLM-as-Judge: исчезновение, искажение и протокол валидации https://korshunov.ai/ko/article/12344-habseong-llm-as-judge-kopeoseuyi-teseuteu-orakeul-munje-sosil-waegog-mic/korshunov.aiko2026-07-16T06:34:26.265387+00:00합성 LLM-as-Judge 코퍼스의 테스트 오라클 문제: 소실, 왜곡 및 검증 프로토콜 https://korshunov.ai/id/article/12344-masalah-oracle-uji-dalam-korpus-sintetis-llm-as-judge-hilang-distorsi-dan/korshunov.aiid2026-07-16T06:34:26.265387+00:00Masalah Oracle Uji dalam Korpus Sintetis LLM-as-Judge: Hilang, Distorsi, dan Protokol Validasi https://korshunov.ai/fr/article/12344-probleme-de-l-oracle-de-test-dans-les-corpus-synthetiques-llm-as-judge-et-un-de/korshunov.aifr2026-07-16T06:34:26.265387+00:00Problème de l'oracle de test dans les corpus synthétiques LLM-as-Judge : Disparition, Distorsion et un protocole de validation https://korshunov.ai/zh-hans/article/12344-he-cheng-llm-as-judgeyu-liao-ku-zhong-de-ce-shi-yu-yan-ji-wen-ti-xiao-shi-niu-qu/korshunov.aizh-cn2026-07-16T06:34:26.265387+00:00合成LLM-as-Judge语料库中的测试预言机问题：消失、扭曲与验证协议 https://korshunov.ai/ar/article/12344-mshkl-lmrj-lkhtbry-fy-mjmw-t-bynt-llm-as-judge-ltrkyby-lkhtf-ltshwyh-wbrwtwkwl/korshunov.aiar2026-07-16T06:34:26.265387+00:00مشكلة المرجع الاختباري في مجموعات بيانات LLM-as-Judge التركيبية: الاختفاء، التشويه وبروتوكول التحقق https://korshunov.ai/es/article/12344-problema-del-oraculo-de-prueba-en-corpora-sinteticos-de-llm-as-judge-distorsion/korshunov.aies2026-07-16T06:34:26.265387+00:00Problema del oráculo de prueba en corpora sintéticos de LLM-as-Judge: desaparición, distorsión y un protocolo de validación https://korshunov.ai/en/article/12344-test-oracle-problem-in-synthetic-llm-as-judge-corpora-disappearance-distortion-a/korshunov.aien2026-07-16T06:34:26.265387+00:00Test Oracle Problem in Synthetic LLM-as-Judge Corpora: Disappearance, Distortion and a Validation Protocol https://korshunov.ai/ko/article/12343-groc-poneun-dajung-modal-hwangageul-juligi-wihan-geungeo-giban-munmaeg-seonho/korshunov.aiko2026-07-16T06:24:32.855161+00:00Groc-PO는 다중 모달 환각을 줄이기 위한 근거 기반 문맥 선호 최적화를 도입 https://korshunov.ai/fr/article/12343-groc-po-introduit-l-optimisation-des-preferences-par-contexte-ancre-pour-reduire/korshunov.aifr2026-07-16T06:24:32.855161+00:00Groc-PO introduit l'optimisation des préférences par contexte ancré pour réduire les hallucinations multimodales https://korshunov.ai/es/article/12343-groc-po-introduce-la-optimizacion-de-preferencia-con-contexto-fundamentado-para/korshunov.aies2026-07-16T06:24:32.855161+00:00Groc-PO introduce la optimización de preferencia con contexto fundamentado para reducir alucinaciones multimodales https://korshunov.ai/zh-hans/article/12343-groc-poyin-ru-ji-yu-jie-di-shang-xia-wen-de-pian-hao-you-hua-yi-jian-shao-duo-mo/korshunov.aizh-cn2026-07-16T06:24:32.855161+00:00Groc-PO引入基于接地上下文的偏好优化以减少多模态幻觉 https://korshunov.ai/ar/article/12343-yqdm-groc-po-thsyn-tfdylt-lsyq-lmtjdhr-ltqlyl-lhlwst-mt-dd-lwsyt/korshunov.aiar2026-07-16T06:24:32.855161+00:00يقدم Groc-PO تحسين تفضيلات السياق المتجذر لتقليل الهلوسات متعددة الوسائط https://korshunov.ai/id/article/12343-groc-po-memperkenalkan-optimasi-preferensi-konteks-terlandaskan-untuk-mengurangi/korshunov.aiid2026-07-16T06:24:32.855161+00:00Groc-PO memperkenalkan optimasi preferensi konteks terlandaskan untuk mengurangi halusinasi multimodal https://korshunov.ai/ru/article/12343-groc-po-predstavliaet-optimizatsiiu-predpochteniia-na-osnove-zazemlennogo-dlia/korshunov.airu2026-07-16T06:24:32.855161+00:00Groc-PO представляет оптимизацию предпочтения на основе заземлённого контекста для снижения мультимодальных галлюцинаций https://korshunov.ai/en/article/12343-groc-po-introduces-grounded-context-preference-optimization-to-reduce-multimodal/korshunov.aien2026-07-16T06:24:32.855161+00:00Groc-PO introduces grounded context preference optimization to reduce multimodal hallucinations https://korshunov.ai/pt-br/article/12343-groc-po-introduz-otimizacao-de-preferencia-com-contexto-fundamentado-para/korshunov.aipt2026-07-16T06:24:32.855161+00:00Groc-PO introduz otimização de preferência com contexto fundamentado para reduzir alucinações multimodais https://korshunov.ai/ja/article/12343-groc-poha-maruchimo-taruhuan-jue-woqing-jian-surutamenojie-di-wen-mo-you-xian-ru/korshunov.aija2026-07-16T06:24:32.855161+00:00Groc-POは、マルチモーダル幻覚を軽減するための接地文脈優先最適化を導入 https://korshunov.ai/hi/article/12343-groc-po-bhuaayaamii-bhrm-ko-km-krne-ke-lie-aadhaarit-sndrbh-praathmiktaa-pesh/korshunov.aihi2026-07-16T06:24:32.855161+00:00Groc-PO बहुआयामी भ्रम को कम करने के लिए आधारित संदर्भ प्राथमिकता अनुकूलन पेश करता है https://korshunov.ai/fr/article/12342-le-dtw-sur-wavlm-permet-l-evaluation-sans-texte-de-la-parole-l2-pour-la-le-et-l/korshunov.aifr2026-07-16T06:17:34.924970+00:00Le DTW sur WavLM permet l'évaluation sans texte de la parole L2 pour la phonétique, le rythme et l'intonation https://korshunov.ai/ko/article/12342-wavlme-daehan-dtwneun-eumunron-rideum-eogyangeul-wihan-tegseuteu-eobsneun-l2/korshunov.aiko2026-07-16T06:17:34.924970+00:00WavLM에 대한 DTW는 음운론, 리듬, 억양을 위한 텍스트 없는 L2 발화 점수 매기를 가능하게 합니다 https://korshunov.ai/ja/article/12342-wavlmnidui-surudtwniyori-yin-yun-lun-risumu-intone-shiyonnotekisutobu-yao-nadi/korshunov.aija2026-07-16T06:17:34.924970+00:00WavLMに対するDTWにより、音韻論、リズム、イントネーションのテキスト不要な第二言語発話評価が可能に https://korshunov.ai/es/article/12342-dtw-sobre-wavlm-permite-la-puntuacion-de-habla-l2-sin-texto-para-fonetica-ritmo/korshunov.aies2026-07-16T06:17:34.924970+00:00DTW sobre WavLM permite la puntuación de habla L2 sin texto para fonética, ritmo e entonación https://korshunov.ai/zh-hans/article/12342-ji-yu-wavlmde-dtwshi-xian-wu-xu-wen-ben-de-l2yu-yin-ping-fen-han-gai-fa-yin-jie/korshunov.aizh-cn2026-07-16T06:17:34.924970+00:00基于WavLM的DTW实现无需文本的L2语音评分，涵盖发音、节奏和语调 https://korshunov.ai/hi/article/12342-wavlm-pr-dtw-dhvnivijnyaan-ly-aur-svraaroh-ke-lie-paatth-mukt-l2-uccaarnn-skssm/korshunov.aihi2026-07-16T06:17:34.924970+00:00WavLM पर DTW ध्वनिविज्ञान, लय और स्वरारोह के लिए पाठ-मुक्त L2 उच्चारण मूल्यांकन सक्षम बनाता है https://korshunov.ai/en/article/12342-dtw-over-wavlm-enables-text-free-l2-speech-scoring-for-phonetics-rhythm-and/korshunov.aien2026-07-16T06:17:34.924970+00:00DTW over WavLM enables text-free L2 speech scoring for phonetics, rhythm, and intonation https://korshunov.ai/ru/article/12342-dtw-poverkh-wavlm-obespechivaet-otsenku-l2-rechi-bez-teksta-dlia-fonetiki-ritma/korshunov.airu2026-07-16T06:17:34.924970+00:00DTW поверх WavLM обеспечивает оценку L2-речи без текста для фонетики, ритма и интонации https://korshunov.ai/ar/article/12342-tmkyn-tqyym-ntq-llg-lthny-bdwn-ns-llphonetics-wlyq-wlnbr-bstkhdm-dtw-l-wavlm/korshunov.aiar2026-07-16T06:17:34.924970+00:00تمكين تقييم نطق اللغة الثانية بدون نص للphonetics والإيقاع والنبرة باستخدام DTW على WavLM https://korshunov.ai/pt-br/article/12342-dtw-sobre-wavlm-permite-pontuacao-de-fala-l2-sem-texto-para-fonetica-ritmo-e/korshunov.aipt2026-07-16T06:17:34.924970+00:00DTW sobre WavLM permite pontuação de fala L2 sem texto para fonética, ritmo e entonação https://korshunov.ai/id/article/12342-dtw-pada-wavlm-memungkinkan-penilaian-ucapan-l2-tanpa-teks-untuk-fonetik-ritme/korshunov.aiid2026-07-16T06:17:34.924970+00:00DTW pada WavLM memungkinkan penilaian ucapan L2 tanpa teks untuk fonetik, ritme, dan intonasi https://korshunov.ai/hi/article/12380-opencl-knpaailr-smsyaaon-ko-hl-krne-ke-lie-fa-aur-moe-vjn-ke-punh-paiking-ko-hai/korshunov.aihi2026-07-16T06:08:07.959155+00:00OpenCL कंपाइलर समस्याओं को हल करने के लिए FA और MoE वजन के पुनः पैकिंग को अक्षम करता है https://korshunov.ai/fr/article/12380-opencl-desactive-le-reconditionnement-des-poids-fa-et-moe-pour-contourner-les-du/korshunov.aifr2026-07-16T06:08:07.959155+00:00OpenCL désactive le reconditionnement des poids FA et MoE pour contourner les problèmes du compilateur https://korshunov.ai/ko/article/12380-opencl-keompailreo-munje-uhoereul-wihae-fa-mic-moe-gajungci-jaepaeking/korshunov.aiko2026-07-16T06:08:07.959155+00:00OpenCL, 컴파일러 문제 우회를 위해 FA 및 MoE 가중치 재패킹 비활성화 https://korshunov.ai/ja/article/12380-openclkakonhairanowen-ti-wohui-bi-surutamenifatomoenozhong-mizai-hatsukuwowu-hua/korshunov.aija2026-07-16T06:08:07.959155+00:00OpenCLがコンパイラの問題を回避するためにFAとMoEの重み再パックを無効化 https://korshunov.ai/pt-br/article/12380-opencl-desativa-a-reorganizacao-de-pesos-fa-e-moe-para-contornar-problemas-do/korshunov.aipt2026-07-16T06:08:07.959155+00:00OpenCL desativa a reorganização de pesos FA e MoE para contornar problemas do compilador https://korshunov.ai/zh-hans/article/12380-opencl-jin-yong-fa-he-moe-quan-zhong-de-zhong-xin-da-bao-yi-gui-bi-bian-yi-qi-ti/korshunov.aizh-cn2026-07-16T06:08:07.959155+00:00OpenCL 禁用 FA 和 MoE 权重的重新打包以规避编译器问题 https://korshunov.ai/es/article/12380-opencl-desactiva-la-reorganizacion-de-pesos-fa-y-moe-para-evitar-problemas-del/korshunov.aies2026-07-16T06:08:07.959155+00:00OpenCL desactiva la reorganización de pesos FA y MoE para evitar problemas del compilador https://korshunov.ai/ru/article/12380-opencl-otkliuchaet-pereupakovku-vesov-fa-i-moe-dlia-obkhoda-problem-kompiliatora/korshunov.airu2026-07-16T06:08:07.959155+00:00OpenCL отключает переупаковку весов FA и MoE для обхода проблем компилятора https://korshunov.ai/en/article/12380-opencl-disables-fa-and-moe-weights-repack-to-workaround-compiler-issues/korshunov.aien2026-07-16T06:08:07.959155+00:00OpenCL disables FA and MoE weights repack to workaround compiler issues https://korshunov.ai/ar/article/12380-t-tyl-d-t-by-wzn-fa-w-moe-fy-opencl-ltjwz-mshklt-lmtrjm/korshunov.aiar2026-07-16T06:08:07.959155+00:00تعطيل إعادة تعبئة أوزان FA و MoE في OpenCL لتجاوز مشكلات المترجم https://korshunov.ai/id/article/12380-opencl-menonaktifkan-repacking-bobot-fa-dan-moe-untuk-mengatasi-masalah-kompiler/korshunov.aiid2026-07-16T06:08:07.959155+00:00OpenCL menonaktifkan repacking bobot FA dan MoE untuk mengatasi masalah kompiler https://korshunov.ai/pt-br/article/12381-lancamento-do-llama-cpp-b10035-com-correcao-de-extracao-q1-0-para-cuda/korshunov.aipt2026-07-16T06:05:26.929390+00:00Lançamento do llama.cpp b10035 com correção de extração Q1_0 para CUDA https://korshunov.ai/hi/article/12381-cuda-q1-0-eksttraikshn-phiks-ke-saath-llama-cpp-b10035-riliij/korshunov.aihi2026-07-16T06:05:26.929390+00:00CUDA Q1_0 एक्सट्रैक्शन फिक्स के साथ llama.cpp b10035 रिलीज़ https://korshunov.ai/en/article/12381-llama-cpp-b10035-release-with-cuda-q1-0-extraction-fix/korshunov.aien2026-07-16T06:05:26.929390+00:00llama.cpp b10035 release with CUDA Q1_0 extraction fix https://korshunov.ai/ja/article/12381-cuda-q1-0chou-chu-xiu-zheng-wohan-mullama-cpp-b10035riri-su/korshunov.aija2026-07-16T06:05:26.929390+00:00CUDA Q1_0抽出修正を含むllama.cpp b10035リリース https://korshunov.ai/ru/article/12381-reliz-llama-cpp-b10035-s-ispravleniem-izvlecheniia-q1-0-dlia-cuda/korshunov.airu2026-07-16T06:05:26.929390+00:00Релиз llama.cpp b10035 с исправлением извлечения Q1_0 для CUDA https://korshunov.ai/es/article/12381-lanzamiento-de-llama-cpp-b10035-con-correccion-de-extraccion-q1-0-para-cuda/korshunov.aies2026-07-16T06:05:26.929390+00:00Lanzamiento de llama.cpp b10035 con corrección de extracción Q1_0 para CUDA https://korshunov.ai/zh-hans/article/12381-llama-cpp-b10035-fa-bu-xiu-fu-cuda-q1-0-ti-qu-wen-ti/korshunov.aizh-cn2026-07-16T06:05:26.929390+00:00llama.cpp b10035 发布，修复 CUDA Q1_0 提取问题 https://korshunov.ai/ar/article/12381-sdr-llama-cpp-b10035-m-slh-stkhrj-q1-0-lkhs-b-cuda/korshunov.aiar2026-07-16T06:05:26.929390+00:00إصدار llama.cpp b10035 مع إصلاح استخراج Q1_0 الخاص بـ CUDA https://korshunov.ai/id/article/12381-rilis-llama-cpp-b10035-dengan-perbaikan-ekstraksi-q1-0-cuda/korshunov.aiid2026-07-16T06:05:26.929390+00:00Rilis llama.cpp b10035 dengan perbaikan ekstraksi Q1_0 CUDA https://korshunov.ai/fr/article/12381-publication-de-llama-cpp-b10035-avec-correction-d-extraction-q1-0-cuda/korshunov.aifr2026-07-16T06:05:26.929390+00:00Publication de llama.cpp b10035 avec correction d'extraction Q1_0 CUDA https://korshunov.ai/ko/article/12381-cuda-q1-0-cucul-sujeongi-pohamdoen-llama-cpp-b10035-rilriseu/korshunov.aiko2026-07-16T06:05:26.929390+00:00CUDA Q1_0 추출 수정이 포함된 llama.cpp b10035 릴리스 https://korshunov.ai/pt-br/article/12341-analise-em-tres-etapas-mostra-que-sft-rl-e-opd-moldam-a-confianca-no-raciocinio/korshunov.aipt2026-07-16T05:56:39.372667+00:00Análise em três etapas mostra que SFT, RL e OPD moldam a confiança no raciocínio de maneira diferente https://korshunov.ai/ar/article/12341-thlyl-mn-thlth-mrhl-yuzhr-n-sft-wrl-wopd-yshkln-thq-cot-bshkl-mkhtlf/korshunov.aiar2026-07-16T05:56:39.372667+00:00تحليل من ثلاث مراحل يُظهر أن SFT وRL وOPD يشكلان ثقة CoT بشكل مختلف https://korshunov.ai/zh-hans/article/12341-san-jie-duan-fen-xi-xian-shi-sft-rl-he-opd-yi-bu-tong-fang-shi-su-zao-si-wei-xin/korshunov.aizh-cn2026-07-16T05:56:39.372667+00:00三阶段分析显示 SFT、RL 和 OPD 以不同方式塑造思维链信心 https://korshunov.ai/id/article/12341-analisis-tiga-tahap-menunjukkan-sft-rl-dan-opd-membentuk-kepercayaan-cot-secara/korshunov.aiid2026-07-16T05:56:39.372667+00:00Analisis tiga tahap menunjukkan SFT, RL, dan OPD membentuk kepercayaan CoT secara berbeda https://korshunov.ai/hi/article/12341-tiin-crnn-vishlessnn-dikhaataa-hai-ki-sft-rl-aur-opd-cot-aatmvishvaas-ko-alg-alg/korshunov.aihi2026-07-16T05:56:39.372667+00:00तीन-चरण विश्लेषण दिखाता है कि SFT, RL और OPD CoT आत्मविश्वास को अलग-अलग ढालते हैं https://korshunov.ai/fr/article/12341-une-analyse-en-trois-etapes-montre-que-le-sft-le-rl-et-l-opd-faconnent-la-du-cot/korshunov.aifr2026-07-16T05:56:39.372667+00:00Une analyse en trois étapes montre que le SFT, le RL et l'OPD façonnent différemment la confiance du CoT https://korshunov.ai/es/article/12341-un-analisis-en-tres-etapas-muestra-que-sft-rl-y-opd-moldean-la-confianza-en-el/korshunov.aies2026-07-16T05:56:39.372667+00:00Un análisis en tres etapas muestra que SFT, RL y OPD moldean la confianza en el razonamiento de manera diferente https://korshunov.ai/ru/article/12341-trekhetapnyi-analiz-pokazyvaet-chto-sft-rl-i-opd-po-raznomu-formiruiut-v/korshunov.airu2026-07-16T05:56:39.372667+00:00Трёхэтапный анализ показывает, что SFT, RL и OPD по-разному формируют уверенность в рассуждениях https://korshunov.ai/en/article/12341-three-stage-analysis-shows-sft-rl-and-opd-shape-cot-confidence-differently/korshunov.aien2026-07-16T05:56:39.372667+00:00Three-stage analysis shows SFT, RL, and OPD shape CoT confidence differently https://korshunov.ai/ko/article/12341-3dangye-bunseogeun-sft-rl-opdga-cot-sinroedoe-seoro-dareun-yeonghyangeul-geoseul/korshunov.aiko2026-07-16T05:56:39.372667+00:003단계 분석은 SFT, RL, OPD가 CoT 신뢰도에 서로 다른 영향을 미친다는 것을 보여줍니다 https://korshunov.ai/ja/article/12341-san-duan-jie-fen-xi-niyori-sft-rl-opdkacotnoxin-lai-xing-niyi-naruying-xiang/korshunov.aija2026-07-16T05:56:39.372667+00:00三段階分析により、SFT、RL、OPDがCoTの信頼性に異なる影響を与えることが示される https://korshunov.ai/es/article/12340-el-modelo-alice-and-bob-unifica-las-caracterizaciones-de-regularidad-al-estilo/korshunov.aies2026-07-16T05:48:14.550495+00:00El modelo Alice and Bob unifica las caracterizaciones de regularidad al estilo Nerode https://korshunov.ai/ko/article/12340-alicewa-bob-modeli-nerode-seutailyi-jeonggyuseong-teugjingeul-tonghab/korshunov.aiko2026-07-16T05:48:14.550495+00:00Alice와 Bob 모델이 Nerode 스타일의 정규성 특징을 통합 https://korshunov.ai/en/article/12340-alice-and-bob-model-unifies-nerode-style-characterizations-of-regularity/korshunov.aien2026-07-16T05:48:14.550495+00:00Alice and Bob model unifies Nerode-style characterizations of regularity https://korshunov.ai/zh-hans/article/12340-alice-and-bob-mo-xing-tong-yi-liao-nerode-feng-ge-de-zheng-ze-xing-ke-hua/korshunov.aizh-cn2026-07-16T05:48:14.550495+00:00Alice and Bob 模型统一了 Nerode 风格的正则性刻画 https://korshunov.ai/pt-br/article/12340-o-modelo-alice-and-bob-unifica-caracterizacoes-de-regularidade-ao-estilo-nerode/korshunov.aipt2026-07-16T05:48:14.550495+00:00O modelo Alice and Bob unifica caracterizações de regularidade ao estilo Nerode https://korshunov.ai/hi/article/12340-alice-and-bob-monddl-niymittaa-ke-nrodde-shailii-vivrnnon-ko-ekjutt-krtaa-hai/korshunov.aihi2026-07-16T05:48:14.550495+00:00Alice and Bob मॉडल नियमितता के नरोडे-शैली विवरणों को एकजुट करता है https://korshunov.ai/fr/article/12340-le-modele-d-alice-et-bob-unifie-les-caracterisations-de-la-regularite-au-style/korshunov.aifr2026-07-16T05:48:14.550495+00:00Le modèle d'Alice et Bob unifie les caractérisations de la régularité au style Nerode https://korshunov.ai/id/article/12340-model-alice-dan-bob-menyatukan-karakterisasi-reguleritas-gaya-nerode/korshunov.aiid2026-07-16T05:48:14.550495+00:00Model Alice dan Bob menyatukan karakterisasi reguleritas gaya Nerode https://korshunov.ai/ar/article/12340-nmwdhj-lys-wbwb-ywhwd-khsys-lntzm-l-tryq-nyrwd/korshunov.aiar2026-07-16T05:48:14.550495+00:00نموذج أليس وبوب يوحّد خصائص الانتظام على طريقة نيرود https://korshunov.ai/ja/article/12340-alicetobobnomoterukanerodeliu-nozheng-ze-xing-note-xing-wotong-yi/korshunov.aija2026-07-16T05:48:14.550495+00:00AliceとBobのモデルがNerode流の正則性の特性を統一 https://korshunov.ai/ru/article/12340-model-alice-and-bob-obediniaet-kharakteristiki-reguliarnosti-v-stile-nerode/korshunov.airu2026-07-16T05:48:14.550495+00:00Модель Alice and Bob объединяет характеристики регулярности в стиле Нероде https://korshunov.ai/ja/article/12339-spycehagui-dao-wozai-li-yong-ke-neng-nasukirunizheng-liu-shi-horishi-togong-jin/korshunov.aija2026-07-16T05:42:35.107916+00:00SPyCEは軌道を再利用可能なスキルに蒸留し、ポリシーと共進化する https://korshunov.ai/id/article/12339-spyce-memadatkan-lintasan-menjadi-keterampilan-yang-dapat-digunakan-kembali-yang/korshunov.aiid2026-07-16T05:42:35.107916+00:00SPyCE memadatkan lintasan menjadi keterampilan yang dapat digunakan kembali yang berevolusi bersama kebijakan https://korshunov.ai/fr/article/12339-spyce-condense-les-trajectoires-en-competences-reutilisables-qui-coevoluent-avec/korshunov.aifr2026-07-16T05:42:35.107916+00:00SPyCE condense les trajectoires en compétences réutilisables qui coévoluent avec la politique https://korshunov.ai/hi/article/12339-spyce-ttrejekttriij-ko-punh-upyog-yogy-kaushl-men-snghnit-krtaa-hai-jo-ponlisii/korshunov.aihi2026-07-16T05:42:35.107916+00:00SPyCE ट्रेजेक्ट्रीज़ को पुनः उपयोग योग्य कौशल में संघनित करता है जो पॉलिसी के साथ सह-विकास करते हैं https://korshunov.ai/ko/article/12339-spyceneun-gwejeogeul-jaesayong-ganeunghan-gisulro-abcughayeo-jeongcaeggwa/korshunov.aiko2026-07-16T05:42:35.107916+00:00SPyCE는 궤적을 재사용 가능한 기술로 압축하여 정책과 공동 진화합니다 https://korshunov.ai/zh-hans/article/12339-spycejiang-gui-ji-ti-lian-wei-ke-yu-ce-lue-xie-tong-jin-hua-de-ke-zhong-yong-ji/korshunov.aizh-cn2026-07-16T05:42:35.107916+00:00SPyCE将轨迹提炼为可与策略协同进化的可重用技能 https://korshunov.ai/ar/article/12339-tqwm-spyce-btkthyf-lmsrt-l-mhrt-qbl-l-d-lstkhdm-tttwr-bltzmn-m-lsys/korshunov.aiar2026-07-16T05:42:35.107916+00:00تقوم SPyCE بتكثيف المسارات إلى مهارات قابلة لإعادة الاستخدام تتطور بالتزامن مع السياسة https://korshunov.ai/en/article/12339-spyce-distills-trajectories-into-reusable-skills-that-co-evolve-with-the-policy/korshunov.aien2026-07-16T05:42:35.107916+00:00SPyCE distills trajectories into reusable skills that co-evolve with the policy https://korshunov.ai/ru/article/12339-spyce-distilliruet-traektorii-v-pereispolzuemye-navyki-kotorye-sovmestno-s/korshunov.airu2026-07-16T05:42:35.107916+00:00SPyCE дистиллирует траектории в переиспользуемые навыки, которые совместно эволюционируют с политикой https://korshunov.ai/es/article/12339-spyce-destila-trayectorias-en-habilidades-reutilizables-que-coevolucionan-con-la/korshunov.aies2026-07-16T05:42:35.107916+00:00SPyCE destila trayectorias en habilidades reutilizables que coevolucionan con la política https://korshunov.ai/pt-br/article/12339-spyce-destila-trajetorias-em-habilidades-reutilizaveis-que-coevoluem-com-a/korshunov.aipt2026-07-16T05:42:35.107916+00:00SPyCE destila trajetórias em habilidades reutilizáveis que coevoluem com a política https://korshunov.ai/zh-hans/article/12338-duo-yu-yan-yan-jiu-bi-jiao-claim-evidence-reasoning-he-divergent-questioning-yu/korshunov.aizh-cn2026-07-16T05:33:55.443971+00:00多语言研究比较 Claim-Evidence-Reasoning 和 Divergent Questioning 用于高阶问题生成 https://korshunov.ai/hi/article/12338-bhubhaassii-adhyyn-claim-evidence-reasoning-aur-divergent-questioning-kii-tulnaa/korshunov.aihi2026-07-16T05:33:55.443971+00:00बहुभाषी अध्ययन Claim-Evidence-Reasoning और Divergent Questioning की तुलना उच्च-क्रम प्रश्न उत्पन्न करने के लिए करता है https://korshunov.ai/es/article/12338-estudio-multilingue-compara-claim-evidence-reasoning-y-divergent-questioning-la/korshunov.aies2026-07-16T05:33:55.443971+00:00Estudio multilingüe compara Claim-Evidence-Reasoning y Divergent Questioning para la generación de preguntas de alto orden https://korshunov.ai/ru/article/12338-mezhdunarodnoe-issledovanie-sravnivaet-claim-evidence-reasoning-i-divergent-dlia/korshunov.airu2026-07-16T05:33:55.443971+00:00Международное исследование сравнивает Claim-Evidence-Reasoning и Divergent Questioning для генерации вопросов высокого порядка https://korshunov.ai/fr/article/12338-une-etude-multilingue-compare-claim-evidence-reasoning-et-divergent-questioning/korshunov.aifr2026-07-16T05:33:55.443971+00:00Une étude multilingue compare Claim-Evidence-Reasoning et Divergent Questioning pour la génération de questions d'ordre supérieur https://korshunov.ai/ko/article/12338-jujang-jeunggeo-curon-mic-balsanjeog-jilmunhwa-peureimweokeureul-hwalyonghan/korshunov.aiko2026-07-16T05:33:55.443971+00:00주장-증거-추론 및 발산적 질문화 프레임워크를 활용한 고차원 질문 생성에 대한 다국어 연구 https://korshunov.ai/id/article/12338-studi-multibahasa-membandingkan-claim-evidence-reasoning-dan-divergent-untuk/korshunov.aiid2026-07-16T05:33:55.443971+00:00Studi multibahasa membandingkan Claim-Evidence-Reasoning dan Divergent Questioning untuk generasi pertanyaan tingkat tinggi https://korshunov.ai/ar/article/12338-drs-mt-dd-llgt-tqrn-byn-tr-ld-ldlyl-lstdll-wtrh-lsyl-lmtbyn-ltwlyd-syl-ly-lmstw/korshunov.aiar2026-07-16T05:33:55.443971+00:00دراسة متعددة اللغات تقارن بين إطار الادعاء-الدليل-الاستدلال وطرح الأسئلة المتباينة لتوليد أسئلة عالية المستوى https://korshunov.ai/en/article/12338-multilingual-study-compares-claim-evidence-reasoning-and-divergent-questioning/korshunov.aien2026-07-16T05:33:55.443971+00:00Multilingual study compares Claim-Evidence-Reasoning and Divergent Questioning for high-order question generation https://korshunov.ai/pt-br/article/12338-estudo-multilingue-compara-claim-evidence-reasoning-e-divergent-questioning-para/korshunov.aipt2026-07-16T05:33:55.443971+00:00Estudo multilíngue compara Claim-Evidence-Reasoning e Divergent Questioning para geração de perguntas de alto nível https://korshunov.ai/ja/article/12338-kure-mu-ehitensu-reso-ninkutotaiha-shiento-kuesuchiyoninkunobi-jiao-niyoruduo-yu/korshunov.aija2026-07-16T05:33:55.443971+00:00クレーム・エビデンス・レゾーニングとダイバージェント・クエスチョニングの比較による多言語研究 https://korshunov.ai/fr/article/12337-deepstress-teste-sous-stress-les-agents-de-recherche-profonde-face-a-des-preuves/korshunov.aifr2026-07-16T05:27:37.156802+00:00DeepStress teste sous stress les agents de recherche profonde face à des preuves non fiables https://korshunov.ai/ja/article/12337-deepstresskaxin-lai-tekinaizheng-ju-nidui-shitetei-husa-chie/korshunov.aija2026-07-16T05:27:37.156802+00:00DeepStressが信頼できない証拠に対してディープサーチエージェントをストレステスト https://korshunov.ai/en/article/12337-deepstress-stress-tests-deep-search-agents-against-unreliable-evidence/korshunov.aien2026-07-16T05:27:37.156802+00:00DeepStress stress-tests deep search agents against unreliable evidence https://korshunov.ai/pt-br/article/12337-deepstress-testa-agentes-de-busca-profunda-contra-evidencias-nao-confiaveis/korshunov.aipt2026-07-16T05:27:37.156802+00:00DeepStress testa agentes de busca profunda contra evidências não confiáveis https://korshunov.ai/hi/article/12337-deepstress-avishvsniiy-prmaannon-ke-khilaaph-ghn-khoj-ejentts-kaa-sttres-ttestt/korshunov.aihi2026-07-16T05:27:37.156802+00:00DeepStress अविश्वसनीय प्रमाणों के खिलाफ गहन खोज एजेंट्स का स्ट्रेस-टेस्ट करता है https://korshunov.ai/ru/article/12337-deepstress-testiruet-glubinnykh-poiskovykh-agentov-na-ustoichivost-k-nenadezhnym/korshunov.airu2026-07-16T05:27:37.156802+00:00DeepStress тестирует глубинных поисковых агентов на устойчивость к ненадежным данным https://korshunov.ai/ko/article/12337-deepstress-sinroehal-su-eobsneun-jeunggeoe-daehan-dib-seoci-eijeonteu-teseuteu/korshunov.aiko2026-07-16T05:27:37.156802+00:00DeepStress, 신뢰할 수 없는 증거에 대한 딥 서치 에이전트 스트레스 테스트 https://korshunov.ai/zh-hans/article/12337-deepstress-zhen-dui-bu-ke-kao-zheng-ju-dui-shen-du-sou-suo-dai-li-jin-xing-ya-li/korshunov.aizh-cn2026-07-16T05:27:37.156802+00:00DeepStress 针对不可靠证据对深度搜索代理进行压力测试 https://korshunov.ai/es/article/12337-deepstress-somete-a-prueba-de-estres-a-agentes-de-busqueda-profunda-frente-a-no/korshunov.aies2026-07-16T05:27:37.156802+00:00DeepStress somete a prueba de estrés a agentes de búsqueda profunda frente a evidencia no confiable https://korshunov.ai/ar/article/12337-deepstress-ykhtbr-wkl-lbhth-l-myq-tht-ldgt-dd-dl-gyr-mwthwq/korshunov.aiar2026-07-16T05:27:37.156802+00:00DeepStress يختبر وكلاء البحث العميق تحت الضغط ضد أدلة غير موثوقة https://korshunov.ai/id/article/12337-deepstress-menguji-stres-agen-pencarian-mendalam-terhadap-bukti-yang-tidak-dapat/korshunov.aiid2026-07-16T05:27:37.156802+00:00DeepStress menguji stres agen pencarian mendalam terhadap bukti yang tidak dapat diandalkan https://korshunov.ai/ar/article/12336-ywlf-deltamerge-lowres-dltt-llg-wlmhm-lltkyf-m-lmwrd-lmnkhfd/korshunov.aiar2026-07-16T05:20:03.652840+00:00يؤلف DeltaMerge-LowRes دلتات اللغة والمهمة للتكيف مع الموارد المنخفضة https://korshunov.ai/ja/article/12336-deltamerge-lowreshayan-yu-totasukunoterutawozu-mihe-wase-di-zi-yuan-shi-ying/korshunov.aija2026-07-16T05:20:03.652840+00:00DeltaMerge-LowResは言語とタスクのデルタを組み合わせ、低資源適応を実現 https://korshunov.ai/ko/article/12336-deltamerge-lowresneun-eoneo-mic-jageob-deltareul-johabhayeo-jeojaweon/korshunov.aiko2026-07-16T05:20:03.652840+00:00DeltaMerge-LowRes는 언어 및 작업 델타를 조합하여 저자원 적응을 수행합니다 https://korshunov.ai/fr/article/12336-deltamerge-lowres-compose-des-deltas-de-langue-et-de-tache-pour-l-adaptation-aux/korshunov.aifr2026-07-16T05:20:03.652840+00:00DeltaMerge-LowRes compose des deltas de langue et de tâche pour l'adaptation aux ressources faibles https://korshunov.ai/id/article/12336-deltamerge-lowres-menggabungkan-delta-bahasa-dan-tugas-untuk-adaptasi-sumber/korshunov.aiid2026-07-16T05:20:03.652840+00:00DeltaMerge-LowRes menggabungkan delta bahasa dan tugas untuk adaptasi sumber daya rendah https://korshunov.ai/en/article/12336-deltamerge-lowres-composes-language-and-task-deltas-for-low-resource-adaptation/korshunov.aien2026-07-16T05:20:03.652840+00:00DeltaMerge-LowRes composes language and task deltas for low-resource adaptation https://korshunov.ai/ru/article/12336-deltamerge-lowres-kombiniruet-iazykovye-i-zadachnye-delty-dlia-adaptatsii-k/korshunov.airu2026-07-16T05:20:03.652840+00:00DeltaMerge-LowRes комбинирует языковые и задачные дельты для адаптации к ресурсоёмким сценариям https://korshunov.ai/pt-br/article/12336-deltamerge-lowres-compoe-deltas-de-idioma-e-tarefa-para-adaptacao-em-baixa/korshunov.aipt2026-07-16T05:20:03.652840+00:00DeltaMerge-LowRes compõe deltas de idioma e tarefa para adaptação em baixa recursos https://korshunov.ai/hi/article/12336-deltamerge-lowres-km-snsaadhn-anukuuln-ke-lie-bhaassaa-aur-kaary-ddelttaa-ko-hai/korshunov.aihi2026-07-16T05:20:03.652840+00:00DeltaMerge-LowRes कम संसाधन अनुकूलन के लिए भाषा और कार्य डेल्टा को संयोजित करता है https://korshunov.ai/es/article/12336-deltamerge-lowres-compone-deltas-de-idioma-y-tarea-para-adaptacion-en-recursos/korshunov.aies2026-07-16T05:20:03.652840+00:00DeltaMerge-LowRes compone deltas de idioma y tarea para adaptación en recursos limitados https://korshunov.ai/zh-hans/article/12336-deltamerge-lowres-zu-he-yu-yan-he-ren-wu-zeng-liang-yi-shi-xian-di-zi-yuan-gua/korshunov.aizh-cn2026-07-16T05:20:03.652840+00:00DeltaMerge-LowRes 组合语言和任务增量以实现低资源适应 https://korshunov.ai/zh-hans/article/12335-cave-absa-kuang-jia-sheng-cheng-jing-guo-yan-zheng-de-fang-mian-ji-fan-shi-shi/korshunov.aizh-cn2026-07-16T05:12:20.606411+00:00CAVE-ABSA 框架生成经过验证的方面级反事实数据 https://korshunov.ai/pt-br/article/12335-framework-cave-absa-gera-contrafactuais-validados-em-nivel-de-aspecto/korshunov.aipt2026-07-16T05:12:20.606411+00:00Framework CAVE-ABSA gera contrafactuais validados em nível de aspecto https://korshunov.ai/ja/article/12335-cave-absahure-muwa-kukajian-zheng-saretaasuhekutoreherunofan-shi-jia-xiang-cheng/korshunov.aija2026-07-16T05:12:20.606411+00:00CAVE-ABSAフレームワークが検証されたアスペクトレベルの反実仮想を生成 https://korshunov.ai/en/article/12335-cave-absa-framework-generates-validated-aspect-level-counterfactuals/korshunov.aien2026-07-16T05:12:20.606411+00:00CAVE-ABSA framework generates validated aspect-level counterfactuals https://korshunov.ai/ru/article/12335-freimvork-cave-absa-generiruet-proverennye-kontrfakty-na-urovne-aspektov/korshunov.airu2026-07-16T05:12:20.606411+00:00Фреймворк CAVE-ABSA генерирует проверенные контрфакты на уровне аспектов https://korshunov.ai/hi/article/12335-cave-absa-phremvrk-styaapit-phluu-str-ke-kaaunttrphaikcuals-utpnn-krtaa-hai/korshunov.aihi2026-07-16T05:12:20.606411+00:00CAVE-ABSA फ्रेमवर्क सत्यापित पहलू-स्तर के काउंटरफैक्चुअल्स उत्पन्न करता है https://korshunov.ai/ar/article/12335-tr-ml-cave-absa-ywld-mddt-wq-y-counterfactuals-l-mstw-ljwnb-tm-lthqq-mnh/korshunov.aiar2026-07-16T05:12:20.606411+00:00إطار عمل CAVE-ABSA يولد مضادات واقعية (counterfactuals) على مستوى الجوانب تم التحقق منها https://korshunov.ai/id/article/12335-kerangka-kerja-cave-absa-menghasilkan-counterfactual-tingkat-aspek-yang/korshunov.aiid2026-07-16T05:12:20.606411+00:00Kerangka kerja CAVE-ABSA menghasilkan counterfactual tingkat aspek yang tervalidasi https://korshunov.ai/fr/article/12335-le-framework-cave-absa-genere-des-contre-factuels-valides-au-niveau-des-aspects/korshunov.aifr2026-07-16T05:12:20.606411+00:00Le framework CAVE-ABSA génère des contre-factuels validés au niveau des aspects https://korshunov.ai/es/article/12335-el-marco-cave-absa-genera-contrafactuales-validados-a-nivel-de-aspecto/korshunov.aies2026-07-16T05:12:20.606411+00:00El marco CAVE-ABSA genera contrafactuales validados a nivel de aspecto https://korshunov.ai/ko/article/12335-cave-absa-peureimweokeuga-geomjeungdoen-eoseupegteu-sujunyi-bansiljaejeog/korshunov.aiko2026-07-16T05:12:20.606411+00:00CAVE-ABSA 프레임워크가 검증된 어스펙트 수준의 반실재적 가설을 생성 https://korshunov.ai/id/article/12334-relai-vcl-menggandakan-keuntungan-agen-pengoptimal-melalui-pembelajaran/korshunov.aiid2026-07-16T05:04:34.118542+00:00RELAI-VCL menggandakan keuntungan agen-pengoptimal melalui pembelajaran berkelanjutan https://korshunov.ai/en/article/12334-relai-vcl-compounds-agent-optimizer-gains-via-continual-learning/korshunov.aien2026-07-16T05:04:34.118542+00:00RELAI-VCL compounds agent-optimizer gains via continual learning https://korshunov.ai/ar/article/12334-tud-f-relai-vcl-mksb-lwkyl-lmuhswn-br-lt-lm-lmstmr/korshunov.aiar2026-07-16T05:04:34.118542+00:00تُضاعف RELAI-VCL مكاسب الوكيل-المُحسّن عبر التعلم المستمر https://korshunov.ai/zh-hans/article/12334-relai-vcl-tong-guo-chi-xu-xue-xi-shi-dai-li-you-hua-qi-shou-yi-fu-he/korshunov.aizh-cn2026-07-16T05:04:34.118542+00:00RELAI-VCL 通过持续学习使代理-优化器收益复合 https://korshunov.ai/pt-br/article/12334-relai-vcl-compoe-ganhos-de-agente-otimizador-via-aprendizado-continuo/korshunov.aipt2026-07-16T05:04:34.118542+00:00RELAI-VCL compõe ganhos de agente-otimizador via aprendizado contínuo https://korshunov.ai/ja/article/12334-relai-vclhaji-sok-de-xue-xi-wotong-shitee-shientotoohuteimaisa-noxiang-cheng-guo/korshunov.aija2026-07-16T05:04:34.118542+00:00RELAI-VCLは継続的学習を通じてエージェントとオプティマイザーの相乗効果を強化する https://korshunov.ai/hi/article/12334-relai-vcl-nirntr-shikssaa-ke-maadhym-se-ejentt-onpttimaaijr-laabhon-ko-snyukt/korshunov.aihi2026-07-16T05:04:34.118542+00:00RELAI-VCL निरंतर शिक्षा के माध्यम से एजेंट-ऑप्टिमाइज़र लाभों को संयुक्त करता है https://korshunov.ai/ru/article/12334-relai-vcl-usilivaet-preimushchestva-agenta-optimizatora-za-schet-nepreryvnogo/korshunov.airu2026-07-16T05:04:34.118542+00:00RELAI-VCL усиливает преимущества агента-оптимизатора за счёт непрерывного обучения https://korshunov.ai/es/article/12334-relai-vcl-compuesta-ganancias-de-agente-optimizador-mediante-aprendizaje/korshunov.aies2026-07-16T05:04:34.118542+00:00RELAI-VCL compuesta ganancias de agente-optimizador mediante aprendizaje continuo https://korshunov.ai/ko/article/12334-relai-vcleun-jisogjeog-hagseubeul-tonghae-eijeonteu-coejeoghwagi-ijeomeul/korshunov.aiko2026-07-16T05:04:34.118542+00:00RELAI-VCL은 지속적 학습을 통해 에이전트-최적화기 이점을 증폭시킵니다 https://korshunov.ai/fr/article/12334-relai-vcl-amplifie-les-gains-agent-optimiseur-par-l-apprentissage-continu/korshunov.aifr2026-07-16T05:04:34.118542+00:00RELAI-VCL amplifie les gains agent-optimiseur par l'apprentissage continu https://korshunov.ai/id/article/12333-pat-menggunakan-rag-dan-korpus-sebanding-untuk-terjemahan-llm-dokumen-utuh/korshunov.aiid2026-07-16T04:57:14.443805+00:00PAT menggunakan RAG dan korpus sebanding untuk terjemahan LLM dokumen utuh https://korshunov.ai/ar/article/12333-ystkhdm-pat-l-rag-wmjmw-t-lbynt-lqbl-llmqrn-ltrjm-lnmdhj-llgwy-lkbyr-llmstnd/korshunov.aiar2026-07-16T04:57:14.443805+00:00يستخدم PAT الـ RAG ومجموعات البيانات القابلة للمقارنة لترجمة النماذج اللغوية الكبيرة للمستند بأكمله https://korshunov.ai/fr/article/12333-pat-utilise-le-rag-et-des-corpus-comparables-pour-la-traduction-llm-de-documents/korshunov.aifr2026-07-16T04:57:14.443805+00:00PAT utilise le RAG et des corpus comparables pour la traduction LLM de documents entiers https://korshunov.ai/hi/article/12333-pat-puure-dstaavej-llm-anuvaad-ke-lie-rag-aur-tulniiy-kaanpraa-kaa-upyog-krtaa/korshunov.aihi2026-07-16T04:57:14.443805+00:00PAT पूरे-दस्तावेज़ LLM अनुवाद के लिए RAG और तुलनीय कांपरा का उपयोग करता है https://korshunov.ai/pt-br/article/12333-pat-usa-rag-e-corpora-comparaveis-para-traducao-de-documentos-inteiros-por-llm/korshunov.aipt2026-07-16T04:57:14.443805+00:00PAT usa RAG e corpora comparáveis para tradução de documentos inteiros por LLM https://korshunov.ai/ru/article/12333-pat-ispolzuet-rag-i-sopostavimye-korpusy-dlia-perevoda-tselykh-dokumentov-llm/korshunov.airu2026-07-16T04:57:14.443805+00:00PAT использует RAG и сопоставимые корпусы для перевода целых документов LLM https://korshunov.ai/es/article/12333-pat-utiliza-rag-y-corpus-comparables-para-la-traduccion-de-documentos-completos/korshunov.aies2026-07-16T04:57:14.443805+00:00PAT utiliza RAG y corpus comparables para la traducción de documentos completos con LLM https://korshunov.ai/ja/article/12333-patharagtolei-si-ko-hasuwoyong-itawen-shu-quan-ti-noda-gui-mo-yan-yu-moterufan/korshunov.aija2026-07-16T04:57:14.443805+00:00PATはRAGと類似コーパスを用いた文書全体の大規模言語モデル翻訳 https://korshunov.ai/zh-hans/article/12333-pat-shi-yong-rag-he-ke-bi-yu-liao-ku-jin-xing-da-yu-yan-mo-xing-de-zheng-pian-yi/korshunov.aizh-cn2026-07-16T04:57:14.443805+00:00PAT 使用 RAG 和可比语料库进行大语言模型的整篇文档翻译 https://korshunov.ai/en/article/12333-pat-uses-rag-and-comparable-corpora-for-whole-document-llm-translation/korshunov.aien2026-07-16T04:57:14.443805+00:00PAT uses RAG and comparable corpora for whole-document LLM translation https://korshunov.ai/ko/article/12333-pateun-ragwa-bigyo-ganeung-kopeoseureul-hwalyonghan-munseo-jeonce-llm-beonyeog/korshunov.aiko2026-07-16T04:57:14.443805+00:00PAT은 RAG와 비교 가능 코퍼스를 활용한 문서 전체 LLM 번역 https://korshunov.ai/es/article/12332-hindcast-cierra-la-fuga-de-datos-en-la-evaluacion-de-pronosticadores-llm/korshunov.aies2026-07-16T04:48:28.652091+00:00Hindcast cierra la fuga de datos en la evaluación de pronosticadores LLM https://korshunov.ai/zh-hans/article/12332-hindcast-guan-bi-liao-llm-yu-ce-qi-ping-gu-zhong-de-shu-ju-xie-lu/korshunov.aizh-cn2026-07-16T04:48:28.652091+00:00Hindcast 关闭了 LLM 预测器评估中的数据泄露 https://korshunov.ai/en/article/12332-hindcast-closes-data-leakage-in-llm-forecaster-evaluation/korshunov.aien2026-07-16T04:48:28.652091+00:00Hindcast closes data leakage in LLM forecaster evaluation https://korshunov.ai/ar/article/12332-tr-hindcast-yglq-tsrb-lbynt-fy-tqyym-mtnbyy-lnmdhj-llgwy-lkbyr/korshunov.aiar2026-07-16T04:48:28.652091+00:00إطار Hindcast يغلق تسرب البيانات في تقييم متنبئي النماذج اللغوية الكبيرة https://korshunov.ai/id/article/12332-hindcast-menutup-kebocoran-data-dalam-evaluasi-peramal-llm/korshunov.aiid2026-07-16T04:48:28.652091+00:00Hindcast menutup kebocoran data dalam evaluasi peramal LLM https://korshunov.ai/ru/article/12332-hindcast-ustraniaet-utechku-dannykh-pri-otsenke-prognozov-llm/korshunov.airu2026-07-16T04:48:28.652091+00:00Hindcast устраняет утечку данных при оценке прогнозов LLM https://korshunov.ai/fr/article/12332-hindcast-comble-les-fuites-de-donnees-dans-l-evaluation-des-previsionnistes-llm/korshunov.aifr2026-07-16T04:48:28.652091+00:00Hindcast comble les fuites de données dans l'évaluation des prévisionnistes LLM https://korshunov.ai/ja/article/12332-hindcastkallmyu-ce-ping-jia-niokerute-tari-ke-shiwojie-xiao/korshunov.aija2026-07-16T04:48:28.652091+00:00HindcastがLLM予測評価におけるデータリーケージを解消 https://korshunov.ai/ko/article/12332-hindcastga-llm-yeceug-pyeonggayi-deiteo-yuculeul-cadan/korshunov.aiko2026-07-16T04:48:28.652091+00:00Hindcast가 LLM 예측 평가의 데이터 유출을 차단 https://korshunov.ai/pt-br/article/12332-hindcast-fecha-vazamento-de-dados-na-avaliacao-de-forecasters-llm/korshunov.aipt2026-07-16T04:48:28.652091+00:00Hindcast fecha vazamento de dados na avaliação de forecasters LLM https://korshunov.ai/hi/article/12332-hindcast-llm-phorkaasttr-muulyaankn-men-ddettaa-liikej-ko-bnd-krtaa-hai/korshunov.aihi2026-07-16T04:48:28.652091+00:00Hindcast LLM फोरकास्टर मूल्यांकन में डेटा लीकेज को बंद करता है https://korshunov.ai/fr/article/12289-la-conception-du-bloc-feedforward-transformer-preserve-le-rang-du-gradient-sur/korshunov.aifr2026-07-16T04:40:56.710895+00:00La conception du bloc feedforward Transformer préserve le rang du gradient sur la profondeur https://korshunov.ai/ko/article/12289-transformer-sunjeonpa-beulrog-seolgyeneun-gipie-geolcyeo-giulgi-raengkeureul/korshunov.aiko2026-07-16T04:40:56.710895+00:00Transformer 순전파 블록 설계는 깊이에 걸쳐 기울기 랭크를 보존한다 https://korshunov.ai/pt-br/article/12289-o-design-do-bloco-feedforward-do-transformer-preserva-a-rank-do-gradiente-ao-da/korshunov.aipt2026-07-16T04:40:56.710895+00:00O design do bloco feedforward do Transformer preserva a rank do gradiente ao longo da profundidade https://korshunov.ai/zh-hans/article/12289-transformerqian-kui-kuai-she-ji-zai-shen-du-fang-xiang-shang-bao-chi-ti-du-zhi/korshunov.aizh-cn2026-07-16T04:40:56.710895+00:00Transformer前馈块设计在深度方向上保持梯度秩 https://korshunov.ai/ja/article/12289-transformernoshun-chuan-bo-hurotsukushe-ji-hashen-saquan-ti-tegou-pei-rankuwobao/korshunov.aija2026-07-16T04:40:56.710895+00:00Transformerの順伝播ブロック設計は深さ全体で勾配ランクを保持する https://korshunov.ai/hi/article/12289-transformer-phiiddphonrvrdd-blonk-ddijaain-ghraaii-bhr-men-greddiyntt-raink-ko/korshunov.aihi2026-07-16T04:40:56.710895+00:00Transformer फ़ीडफॉरवर्ड ब्लॉक डिज़ाइन गहराई भर में ग्रेडियंट रैंक को बनाए रखता है https://korshunov.ai/en/article/12289-transformer-feedforward-block-design-preserves-gradient-rank-across-depth/korshunov.aien2026-07-16T04:40:56.710895+00:00Transformer feedforward block design preserves gradient rank across depth https://korshunov.ai/ru/article/12289-dizain-bloka-priamoi-peredachi-transformera-sokhraniaet-rang-gradienta-po/korshunov.airu2026-07-16T04:40:56.710895+00:00Дизайн блока прямой передачи трансформера сохраняет ранг градиента по глубине https://korshunov.ai/es/article/12289-el-diseno-del-bloque-de-alimentacion-directa-del-transformer-preserva-el-rango-a/korshunov.aies2026-07-16T04:40:56.710895+00:00El diseño del bloque de alimentación directa del Transformer preserva el rango del gradiente a lo largo de la profundidad https://korshunov.ai/ar/article/12289-tsmym-ktl-ltgdhy-lmmy-fy-lmhwwl-yhfz-l-rtb-ltdrj-br-l-mq/korshunov.aiar2026-07-16T04:40:56.710895+00:00تصميم كتلة التغذية الأمامية في المحوّل يحافظ على رتبة التدرج عبر العمق https://korshunov.ai/id/article/12289-desain-blok-feedforward-transformer-mempertahankan-rank-gradien-melintasi/korshunov.aiid2026-07-16T04:40:56.710895+00:00Desain blok feedforward Transformer mempertahankan rank gradien melintasi kedalaman https://korshunov.ai/id/article/12288-csfs-mengurangi-biaya-pemilihan-fitur-prediksi-daya-angin-dan-surya-sebesar-21/korshunov.aiid2026-07-16T04:31:18.841833+00:00CSFS mengurangi biaya pemilihan fitur prediksi daya angin dan surya sebesar 21% https://korshunov.ai/ar/article/12288-csfs-yqll-tklf-khtyr-myzt-ltnbw-btq-lryh-wltq-lshmsy-bnsb-21/korshunov.aiar2026-07-16T04:31:18.841833+00:00CSFS يقلل تكلفة اختيار ميزات التنبؤ بطاقة الرياح والطاقة الشمسية بنسبة 21% https://korshunov.ai/ko/article/12288-csfs-pungryeog-mic-taeyanggwang-baljeon-yeceug-teugjing-seontaeg-biyong-21/korshunov.aiko2026-07-16T04:31:18.841833+00:00CSFS, 풍력 및 태양광 발전 예측 특징 선택 비용 21% 절감 https://korshunov.ai/fr/article/12288-csfs-reduit-de-21-le-cout-de-selection-des-caracteristiques-pour-la-prediction-l/korshunov.aifr2026-07-16T04:31:18.841833+00:00CSFS réduit de 21 % le coût de sélection des caractéristiques pour la prédiction de l'énergie éolienne et solaire https://korshunov.ai/ja/article/12288-csfskafeng-li-tai-yang-guang-fa-dian-noyu-ce-niokerute-zheng-liang-xuan-ze-xue/korshunov.aija2026-07-16T04:31:18.841833+00:00CSFSが風力・太陽光発電の予測における特徴量選択コストを21%削減 https://korshunov.ai/hi/article/12288-csfs-se-hvaa-aur-saur-uurjaa-bhvissyvaannii-ke-lie-phiicr-cyn-kii-laagt-men-21/korshunov.aihi2026-07-16T04:31:18.841833+00:00CSFS से हवा और सौर ऊर्जा भविष्यवाणी के लिए फीचर चयन की लागत में 21% की कमी https://korshunov.ai/en/article/12288-csfs-reduces-wind-and-solar-power-prediction-feature-selection-cost-by-21/korshunov.aien2026-07-16T04:31:18.841833+00:00CSFS reduces wind and solar power prediction feature selection cost by 21% https://korshunov.ai/ru/article/12288-csfs-snizhaet-stoimost-otbora-priznakov-dlia-prognozirovaniia-vetrovoi-i-energii/korshunov.airu2026-07-16T04:31:18.841833+00:00CSFS снижает стоимость отбора признаков для прогнозирования ветровой и солнечной энергии на 21% https://korshunov.ai/es/article/12288-csfs-reduce-el-costo-de-seleccion-de-caracteristicas-para-la-prediccion-de-y-en/korshunov.aies2026-07-16T04:31:18.841833+00:00CSFS reduce el costo de selección de características para la predicción de energía eólica y solar en un 21% https://korshunov.ai/zh-hans/article/12288-csfsjiang-feng-neng-he-tai-yang-neng-yu-ce-de-te-zheng-xuan-ze-cheng-ben-jiang/korshunov.aizh-cn2026-07-16T04:31:18.841833+00:00CSFS将风能和太阳能预测的特征选择成本降低21% https://korshunov.ai/pt-br/article/12288-csfs-reduz-em-21-o-custo-de-selecao-de-caracteristicas-para-previsao-de-energia/korshunov.aipt2026-07-16T04:31:18.841833+00:00CSFS reduz em 21% o custo de seleção de características para previsão de energia eólica e solar https://korshunov.ai/fr/article/12287-une-etude-revele-que-l-adoption-des-outils-de-codage-agentiques-est-faible-et/korshunov.aifr2026-07-16T04:22:58.318383+00:00Une étude révèle que l'adoption des outils de codage agentiques est faible et concentrée dans les petits projets GitHub https://korshunov.ai/ar/article/12287-drs-tjd-n-tmd-dwt-lbrmj-lwkyly-mnkhfd-wmtrkz-fy-mshry-github-lsgyr/korshunov.aiar2026-07-16T04:22:58.318383+00:00دراسة تجد أن اعتماد أدوات البرمجة الوكيلية منخفض ومتركز في مشاريع GitHub الصغيرة https://korshunov.ai/id/article/12287-studi-menemukan-adopsi-alat-koding-agentic-rendah-dan-terkonsentrasi-di-proyek/korshunov.aiid2026-07-16T04:22:58.318383+00:00Studi menemukan adopsi alat koding agentic rendah dan terkonsentrasi di proyek GitHub kecil https://korshunov.ai/ja/article/12287-e-shientoxing-ko-teinkutsu-runocai-yong-lu-hadi-ku-xiao-gui-mo-zhong-jiu-jie-guo/korshunov.aija2026-07-16T04:22:58.318383+00:00エージェント型コーディングツールの採用率は低く、小規模なGitHubプロジェクトに集中しているという研究結果 https://korshunov.ai/ko/article/12287-eijeonteu-koding-dogu-caetaegryuleun-najgo-sogyumo-github-peurojegteue-issdaneun/korshunov.aiko2026-07-16T04:22:58.318383+00:00에이전트 코딩 도구 채택률은 낮고 소규모 GitHub 프로젝트에 집중되어 있다는 연구 결과 https://korshunov.ai/es/article/12287-el-estudio-encuentra-que-la-adopcion-de-herramientas-de-codificacion-agentica-es/korshunov.aies2026-07-16T04:22:58.318383+00:00El estudio encuentra que la adopción de herramientas de codificación agéntica es baja y está concentrada en proyectos pequeños de GitHub https://korshunov.ai/ru/article/12287-issledovanie-pokazyvaet-chto-vnedrenie-agentnykh-instrumentov-dlia-nizkoe-i-v/korshunov.airu2026-07-16T04:22:58.318383+00:00Исследование показывает, что внедрение агентных инструментов для программирования низкое и сосредоточено в небольших проектах GitHub https://korshunov.ai/zh-hans/article/12287-yan-jiu-fa-xian-zhi-neng-ti-bian-cheng-gong-ju-de-cai-yong-lu-jiao-di-qie-ji-zai/korshunov.aizh-cn2026-07-16T04:22:58.318383+00:00研究发现智能体编程工具的采用率较低，且集中在小型 GitHub 项目中 https://korshunov.ai/pt-br/article/12287-estudo-revela-que-a-adocao-de-ferramentas-de-codificacao-agentica-e-baixa-e-em/korshunov.aipt2026-07-16T04:22:58.318383+00:00Estudo revela que a adoção de ferramentas de codificação agêntica é baixa e concentrada em pequenos projetos do GitHub https://korshunov.ai/hi/article/12287-adhyyn-paataa-hai-ki-ejenttik-kodding-ttuul-kaa-apnaan-km-hai-aur-chotte-github/korshunov.aihi2026-07-16T04:22:58.318383+00:00अध्ययन पाता है कि एजेंटिक कोडिंग टूल का अपनान कम है और छोटे GitHub प्रोजेक्ट्स में केंद्रित है https://korshunov.ai/en/article/12287-study-finds-agentic-coding-tool-adoption-is-low-and-concentrated-in-small-github/korshunov.aien2026-07-16T04:22:58.318383+00:00Study finds agentic coding tool adoption is low and concentrated in small GitHub projects https://korshunov.ai/ja/article/12286-duo-ling-yu-di-riso-suocrnotamenomaruchiekisuha-toru-teinku-man-zhou-yu-shi-li/korshunov.aija2026-07-16T04:12:36.089253+00:00多領域低リソースOCRのためのマルチエキスパートルーティング：満州語事例研究 https://korshunov.ai/ru/article/12286-marshrutizatsiia-s-uchastiem-neskolkikh-ekspertov-dlia-multisubektnogo-ocr-s-na/korshunov.airu2026-07-16T04:12:36.089253+00:00Маршрутизация с участием нескольких экспертов для мультисубъектного OCR с низким уровнем ресурсов: исследование на маньчжурском языке https://korshunov.ai/es/article/12286-enrutamiento-de-multiples-expertos-para-ocr-multirreino-con-pocos-recursos-un-de/korshunov.aies2026-07-16T04:12:36.089253+00:00Enrutamiento de múltiples expertos para OCR multirreino con pocos recursos: un estudio de caso manchu https://korshunov.ai/en/article/12286-multi-expert-routing-for-multi-domain-low-resource-ocr-a-manchu-case-study/korshunov.aien2026-07-16T04:12:36.089253+00:00Multi-Expert Routing for Multi-Domain Low-Resource OCR: A Manchu Case Study https://korshunov.ai/zh-hans/article/12286-duo-zhuan-jia-lu-you-yong-yu-duo-ling-yu-di-zi-yuan-ocr-yi-man-wen-wei-li-de-yan/korshunov.aizh-cn2026-07-16T04:12:36.089253+00:00多专家路由用于多领域低资源OCR：以满文为例的研究 https://korshunov.ai/pt-br/article/12286-roteamento-de-multiplos-especialistas-para-ocr-multidominio-com-poucos-recursos/korshunov.aipt2026-07-16T04:12:36.089253+00:00Roteamento de Múltiplos Especialistas para OCR Multidomínio com Poucos Recursos: Um Estudo de Caso Manchu https://korshunov.ai/ko/article/12286-dajung-domein-jeojaweon-ocreul-wihan-dajung-jeonmunga-rauting-manjueo-sarye/korshunov.aiko2026-07-16T04:12:36.089253+00:00다중 도메인 저자원 OCR을 위한 다중 전문가 라우팅: 만주어 사례 연구 https://korshunov.ai/hi/article/12286-bhu-vishessjny-ruutting-bhu-ddomen-km-snsaadhit-ocr-ke-lie-ek-mncuu-maamlaa/korshunov.aihi2026-07-16T04:12:36.089253+00:00बहु-विशेषज्ञ रूटिंग बहु-डोमेन कम-संसाधित OCR के लिए: एक मंचू मामला अध्ययन https://korshunov.ai/ar/article/12286-twjyh-mt-dd-lkhbr-l-ocr-mnkhfd-lmwrd-mt-dd-lmjlt-drs-hl-lmnshwry/korshunov.aiar2026-07-16T04:12:36.089253+00:00توجيه متعدد الخبراء لـ OCR منخفض الموارد متعدد المجالات: دراسة حالة لمنشورية https://korshunov.ai/id/article/12286-rute-multi-pakar-untuk-ocr-sumber-daya-rendah-multi-domain-studi-kasus-manchu/korshunov.aiid2026-07-16T04:12:36.089253+00:00Rute Multi-Pakar untuk OCR Sumber Daya Rendah Multi-Domain: Studi Kasus Manchu https://korshunov.ai/fr/article/12286-routage-multi-experts-pour-l-ocr-multi-domaines-a-ressources-limitees-une-etude/korshunov.aifr2026-07-16T04:12:36.089253+00:00Routage multi-experts pour l'OCR multi-domaines à ressources limitées : Une étude de cas mandchou https://korshunov.ai/ar/article/12285-tktshf-msht-evo-2-mqwm-lmddt-lmykrwby-fy-lbynt-lmytjynwmy-bdq-ly/korshunov.aiar2026-07-16T03:59:11.800982+00:00تكتشف مسحات Evo 2 مقاومة المضادات الميكروبية في البيانات الميتاجينومية بدقة عالية https://korshunov.ai/id/article/12285-probe-evo-2-mendeteksi-resistensi-antimikroba-dalam-data-metagenomik-dengan/korshunov.aiid2026-07-16T03:59:11.800982+00:00Probe Evo 2 mendeteksi resistensi antimikroba dalam data metagenomik dengan akurasi tinggi https://korshunov.ai/fr/article/12285-les-sondes-evo-2-detectent-la-resistance-aux-antimicrobiens-dans-les-donnees-une/korshunov.aifr2026-07-16T03:59:11.800982+00:00Les sondes Evo 2 détectent la résistance aux antimicrobiens dans les données métagénomiques avec une grande précision https://korshunov.ai/ko/article/12285-evo-2-peurobeuneun-metagenom-deiteoeseo-hangsaengje-naeseongeul-nopeun-gamji/korshunov.aiko2026-07-16T03:59:11.800982+00:00Evo 2 프로브는 메타게놈 데이터에서 항생제 내성을 높은 정확도로 감지 https://korshunov.ai/ja/article/12285-evo-2huro-huhametakenomute-tanioitekang-jun-nai-xing-wogao-jing-du-tejian-chu/korshunov.aija2026-07-16T03:59:11.800982+00:00Evo 2プローブはメタゲノムデータにおいて抗菌耐性を高精度で検出 https://korshunov.ai/hi/article/12285-evo-2-probs-mettaajenomik-ddettaa-men-ucc-sttiiktaa-ke-saath-prtijaivik-prtirodh/korshunov.aihi2026-07-16T03:59:11.800982+00:00Evo 2 प्रोब्स मेटाजेनोमिक डेटा में उच्च सटीकता के साथ प्रतिजैविक प्रतिरोध का पता लगाते हैं https://korshunov.ai/pt-br/article/12285-probes-do-evo-2-detectam-resistencia-antimicrobiana-em-dados-metagenomicos-com/korshunov.aipt2026-07-16T03:59:11.800982+00:00Probes do Evo 2 detectam resistência antimicrobiana em dados metagenômicos com alta precisão https://korshunov.ai/zh-hans/article/12285-evo-2tan-zhen-yi-gao-jing-du-jian-ce-hong-ji-yin-zu-shu-ju-zhong-de-kang-jun-su/korshunov.aizh-cn2026-07-16T03:59:11.800982+00:00Evo 2探针以高精度检测宏基因组数据中的抗菌素耐药性 https://korshunov.ai/es/article/12285-evo-2-detecta-resistencia-antimicrobiana-en-datos-metagenomicos-con-alta/korshunov.aies2026-07-16T03:59:11.800982+00:00Evo 2 detecta resistencia antimicrobiana en datos metagenómicos con alta precisión https://korshunov.ai/ru/article/12285-zondy-evo-2-obnaruzhivaiut-ustoichivost-k-antimikrobnym-preparatam-v-dannykh-s/korshunov.airu2026-07-16T03:59:11.800982+00:00Зонды Evo 2 обнаруживают устойчивость к антимикробным препаратам в метагеномных данных с высокой точностью https://korshunov.ai/en/article/12285-evo-2-probes-detect-antimicrobial-resistance-in-metagenomic-data-with-high/korshunov.aien2026-07-16T03:59:11.800982+00:00Evo 2 probes detect antimicrobial resistance in metagenomic data with high accuracy https://korshunov.ai/ko/article/12284-metaperchneun-bojo-metadeiteoreul-sayonghayeo-baioeumhyang-paundeisyeon-modeleul/korshunov.aiko2026-07-16T03:43:06.943081+00:00MetaPerch는 보조 메타데이터를 사용하여 바이오음향 파운데이션 모델을 개선합니다 https://korshunov.ai/ja/article/12284-metaperchhabu-zhu-metate-tawoshi-yong-shitehaioyin-xiang-ji-pan-moteruwogai-shan/korshunov.aija2026-07-16T03:43:06.943081+00:00MetaPerchは補助メタデータを使用してバイオ音響基盤モデルを改善する https://korshunov.ai/ar/article/12284-ystkhdm-metaperch-bynt-wsfy-ms-d-lthsyn-nmdhj-lss-lswty-lhywy/korshunov.aiar2026-07-16T03:43:06.943081+00:00يستخدم MetaPerch بيانات وصفية مساعدة لتحسين نماذج الأسس الصوتية الحيوية https://korshunov.ai/ru/article/12284-metaperch-ispolzuet-vspomogatelnye-metadannye-dlia-uluchsheniia-fundamentalnykh/korshunov.airu2026-07-16T03:43:06.943081+00:00MetaPerch использует вспомогательные метаданные для улучшения биоакустических фундаментальных моделей https://korshunov.ai/en/article/12284-metaperch-uses-auxiliary-metadata-to-improve-bioacoustic-foundation-models/korshunov.aien2026-07-16T03:43:06.943081+00:00MetaPerch uses auxiliary metadata to improve bioacoustic foundation models https://korshunov.ai/fr/article/12284-metaperch-utilise-des-metadonnees-auxiliaires-pour-ameliorer-les-modeles-de/korshunov.aifr2026-07-16T03:43:06.943081+00:00MetaPerch utilise des métadonnées auxiliaires pour améliorer les modèles de fondation bioacoustiques https://korshunov.ai/zh-hans/article/12284-metaperch-li-yong-fu-zhu-yuan-shu-ju-gai-jin-sheng-wu-sheng-xue-ji-chu-mo-xing/korshunov.aizh-cn2026-07-16T03:43:06.943081+00:00MetaPerch 利用辅助元数据改进生物声学基础模型 https://korshunov.ai/id/article/12284-metaperch-menggunakan-metadata-tambahan-untuk-meningkatkan-model-fondasi/korshunov.aiid2026-07-16T03:43:06.943081+00:00MetaPerch menggunakan metadata tambahan untuk meningkatkan model fondasi bioakustik https://korshunov.ai/es/article/12284-metaperch-utiliza-metadatos-auxiliares-para-mejorar-los-modelos-fundamentales-de/korshunov.aies2026-07-16T03:43:06.943081+00:00MetaPerch utiliza metadatos auxiliares para mejorar los modelos fundamentales de bioacústica https://korshunov.ai/hi/article/12284-metaperch-baayoekausttik-phaaunddeshn-monddl-ko-behtr-bnaane-ke-lie-shaayk-kaa/korshunov.aihi2026-07-16T03:43:06.943081+00:00MetaPerch बायोएकौस्टिक फाउंडेशन मॉडल को बेहतर बनाने के लिए सहायक मेटाडेटा का उपयोग करता है https://korshunov.ai/pt-br/article/12284-metaperch-usa-metadados-auxiliares-para-melhorar-modelos-fundamentais-de/korshunov.aipt2026-07-16T03:43:06.943081+00:00MetaPerch usa metadados auxiliares para melhorar modelos fundamentais de bioacústica https://korshunov.ai/es/article/12283-ot-ica-utiliza-transporte-optimo-para-recuperar-componentes-independientes/korshunov.aies2026-07-16T03:37:35.189454+00:00OT-ICA utiliza transporte óptimo para recuperar componentes independientes https://korshunov.ai/fr/article/12283-ot-ica-utilise-le-transport-optimal-pour-recuperer-les-composants-independants/korshunov.aifr2026-07-16T03:37:35.189454+00:00OT-ICA utilise le transport optimal pour récupérer les composants indépendants https://korshunov.ai/pt-br/article/12283-ot-ica-usa-transporte-otimo-para-recuperar-componentes-independentes/korshunov.aipt2026-07-16T03:37:35.189454+00:00OT-ICA usa transporte ótimo para recuperar componentes independentes https://korshunov.ai/hi/article/12283-ot-ica-svtntr-ghttkon-ko-punrpraapt-krne-ke-lie-isstttm-privhn-kaa-upyog-krtaa/korshunov.aihi2026-07-16T03:37:35.189454+00:00OT-ICA स्वतंत्र घटकों को पुनर्प्राप्त करने के लिए इष्टतम परिवहन का उपयोग करता है https://korshunov.ai/ru/article/12283-ot-ica-ispolzuet-optimalnyi-transport-dlia-vosstanovleniia-nezavisimykh/korshunov.airu2026-07-16T03:37:35.189454+00:00OT-ICA использует оптимальный транспорт для восстановления независимых компонент https://korshunov.ai/zh-hans/article/12283-ot-icali-yong-zui-you-chuan-shu-hui-fu-du-li-fen-liang/korshunov.aizh-cn2026-07-16T03:37:35.189454+00:00OT-ICA利用最优传输恢复独立分量 https://korshunov.ai/en/article/12283-ot-ica-uses-optimal-transport-to-recover-independent-components/korshunov.aien2026-07-16T03:37:35.189454+00:00OT-ICA uses optimal transport to recover independent components https://korshunov.ai/ko/article/12283-ot-icaneun-coejeog-unsongeul-sayonghayeo-dogrib-seongbun-bogweon/korshunov.aiko2026-07-16T03:37:35.189454+00:00OT-ICA는 최적 운송을 사용하여 독립 성분 복원 https://korshunov.ai/ja/article/12283-ot-icahazui-shi-shu-song-woyong-itedu-li-cheng-fen-wofu-yuan-suru/korshunov.aija2026-07-16T03:37:35.189454+00:00OT-ICAは最適輸送を用いて独立成分を復元する https://korshunov.ai/ar/article/12283-ystkhdm-ot-ica-lnql-lmthl-lst-d-lmkwnt-lmstql/korshunov.aiar2026-07-16T03:37:35.189454+00:00يستخدم OT-ICA النقل الأمثل لاستعادة المكونات المستقلة https://korshunov.ai/id/article/12283-ot-ica-menggunakan-transport-optimal-untuk-memulihkan-komponen-independen/korshunov.aiid2026-07-16T03:37:35.189454+00:00OT-ICA menggunakan transport optimal untuk memulihkan komponen independen https://korshunov.ai/ru/article/12282-freimvork-mojo-ispolzuet-obuchenie-bez-uchitelia-dlia-uluchsheniia-neironnykh/korshunov.airu2026-07-16T03:30:51.566666+00:00Фреймворк MOJO использует обучение без учителя для улучшения декодирования нейронных популяций https://korshunov.ai/hi/article/12282-mojo-phremvrk-tntrikaa-aabaadii-ddikodding-ko-behtr-bnaane-ke-lie-sv-suprvaaijdd/korshunov.aihi2026-07-16T03:30:51.566666+00:00MOJO फ्रेमवर्क तंत्रिका आबादी डिकोडिंग को बेहतर बनाने के लिए स्व-सुपरवाइज्ड लर्निंग का उपयोग करता है https://korshunov.ai/pt-br/article/12282-framework-mojo-usa-aprendizado-auto-supervisionado-para-melhorar-a-decodificacao/korshunov.aipt2026-07-16T03:30:51.566666+00:00Framework MOJO usa aprendizado auto-supervisionado para melhorar a decodificação de populações neurais https://korshunov.ai/id/article/12282-kerangka-kerja-mojo-menggunakan-pembelajaran-tanpa-pengawasan-untuk-meningkatkan/korshunov.aiid2026-07-16T03:30:51.566666+00:00Kerangka kerja MOJO menggunakan pembelajaran tanpa pengawasan untuk meningkatkan decoding populasi saraf https://korshunov.ai/ar/article/12282-tr-ml-mojo-ystkhdm-lt-lm-ldhty-lshrfy-lthsyn-fk-tshfyr-lmjmw-t-l-sby/korshunov.aiar2026-07-16T03:30:51.566666+00:00إطار عمل MOJO يستخدم التعلم الذاتي الإشرافي لتحسين فك تشفير المجموعات العصبية https://korshunov.ai/es/article/12282-el-marco-mojo-utiliza-aprendizaje-no-supervisado-para-mejorar-la-decodificacion/korshunov.aies2026-07-16T03:30:51.566666+00:00El marco MOJO utiliza aprendizaje no supervisado para mejorar la decodificación de poblaciones neuronales https://korshunov.ai/zh-hans/article/12282-mojo-kuang-jia-li-yong-zi-jian-du-xue-xi-gai-jin-shen-jing-qun-ti-jie-ma/korshunov.aizh-cn2026-07-16T03:30:51.566666+00:00MOJO 框架利用自监督学习改进神经群体解码 https://korshunov.ai/fr/article/12282-le-framework-mojo-utilise-l-apprentissage-auto-supervise-pour-ameliorer-le-des/korshunov.aifr2026-07-16T03:30:51.566666+00:00Le framework MOJO utilise l'apprentissage auto-supervisé pour améliorer le décodage des populations neuronales https://korshunov.ai/ja/article/12282-mojohure-muwa-kuhazi-ji-jiao-shi-arixue-xi-woyong-iteshen-jing-ji-tuan-noteko/korshunov.aija2026-07-16T03:30:51.566666+00:00MOJOフレームワークは自己教師あり学習を用いて神経集団のデコーディングを改善する https://korshunov.ai/ko/article/12282-mojo-peureimweokeuneun-jagijido-hagseubeul-tonghae-singyeong-jibdan-dikodingeul/korshunov.aiko2026-07-16T03:30:51.566666+00:00MOJO 프레임워크는 자기지도 학습을 통해 신경 집단 디코딩을 개선합니다 https://korshunov.ai/en/article/12282-mojo-framework-uses-self-supervised-learning-to-improve-neural-population/korshunov.aien2026-07-16T03:30:51.566666+00:00MOJO framework uses self-supervised learning to improve neural population decoding https://korshunov.ai/ja/article/12279-haifan-gongkahugging-face-hubhenoctgroundbenchhenchima-kuzhui-jia-worikuesuto/korshunov.aija2026-07-16T03:21:26.495876+00:00Haifan GongがHugging Face HubへのCTGroundBenchベンチマーク追加をリクエスト https://korshunov.ai/en/article/12279-haifan-gong-requests-ctgroundbench-benchmark-addition-to-hugging-face-hub/korshunov.aien2026-07-16T03:21:26.495876+00:00Haifan Gong requests CTGroundBench benchmark addition to Hugging Face Hub https://korshunov.ai/pt-br/article/12279-haifan-gong-solicita-a-adicao-do-benchmark-ctgroundbench-ao-hugging-face-hub/korshunov.aipt2026-07-16T03:21:26.495876+00:00Haifan Gong solicita a adição do benchmark CTGroundBench ao Hugging Face Hub https://korshunov.ai/hi/article/12279-haaiphaan-gong-ne-hugging-face-hub-men-ctgroundbench-bencmaark-joddne-kaa-kiyaa/korshunov.aihi2026-07-16T03:21:26.495876+00:00हाइफान गोंग ने Hugging Face Hub में CTGroundBench बेंचमार्क जोड़ने का अनुरोध किया https://korshunov.ai/ko/article/12279-haifan-gong-hugging-face-hube-ctgroundbench-bencimakeu-cuga-yoceong/korshunov.aiko2026-07-16T03:21:26.495876+00:00Haifan Gong, Hugging Face Hub에 CTGroundBench 벤치마크 추가 요청 https://korshunov.ai/zh-hans/article/12279-haifan-gong-qing-qiu-jiang-ctgroundbench-ji-zhun-tian-jia-dao-hugging-face-hub/korshunov.aizh-cn2026-07-16T03:21:26.495876+00:00Haifan Gong 请求将 CTGroundBench 基准添加到 Hugging Face Hub https://korshunov.ai/ru/article/12279-khaifan-gun-prosit-dobavit-benchmark-ctgroundbench-v-hugging-face-hub/korshunov.airu2026-07-16T03:21:26.495876+00:00Хайфан Гун просит добавить бенчмарк CTGroundBench в Hugging Face Hub https://korshunov.ai/fr/article/12279-haifan-gong-demande-l-ajout-du-benchmark-ctgroundbench-au-hub-hugging-face/korshunov.aifr2026-07-16T03:21:26.495876+00:00Haifan Gong demande l'ajout du benchmark CTGroundBench au Hub Hugging Face https://korshunov.ai/es/article/12279-haifan-gong-solicita-la-adicion-del-benchmark-ctgroundbench-a-hugging-face-hub/korshunov.aies2026-07-16T03:21:26.495876+00:00Haifan Gong solicita la adición del benchmark CTGroundBench a Hugging Face Hub https://korshunov.ai/ar/article/12279-hyfn-gwng-ytlb-df-m-yr-ctgroundbench-l-hugging-face-hub/korshunov.aiar2026-07-16T03:21:26.495876+00:00هايفان غونغ يطلب إضافة معيار CTGroundBench إلى Hugging Face Hub https://korshunov.ai/id/article/12279-haifan-gong-meminta-penambahan-benchmark-ctgroundbench-ke-hugging-face-hub/korshunov.aiid2026-07-16T03:21:26.495876+00:00Haifan Gong meminta penambahan benchmark CTGroundBench ke Hugging Face Hub https://korshunov.ai/es/article/12278-el-usuario-pregunta-como-alojar-conjuntos-de-datos-de-capas-gis-como-enlaces-en/korshunov.aies2026-07-16T03:12:40.722508+00:00El usuario pregunta cómo alojar conjuntos de datos de capas GIS como enlaces HTTP en Hugging Face Spaces https://korshunov.ai/ru/article/12278-polzovatel-sprashivaet-kak-razmeshchat-nabory-dannykh-sloev-gis-v-vide-http-na/korshunov.airu2026-07-16T03:12:40.722508+00:00Пользователь спрашивает, как размещать наборы данных слоёв ГИС в виде HTTP-ссылок на Hugging Face Spaces https://korshunov.ai/fr/article/12278-un-utilisateur-demande-comment-heberger-des-jeux-de-donnees-de-couches-gis-sous/korshunov.aifr2026-07-16T03:12:40.722508+00:00Un utilisateur demande comment héberger des jeux de données de couches GIS sous forme de liens HTTP sur Hugging Face Spaces https://korshunov.ai/pt-br/article/12278-usuario-pergunta-como-hospedar-conjuntos-de-dados-de-camadas-gis-como-links-http/korshunov.aipt2026-07-16T03:12:40.722508+00:00Usuário pergunta como hospedar conjuntos de dados de camadas GIS como links HTTP no Hugging Face Spaces https://korshunov.ai/ko/article/12278-sayongjaga-hugging-face-spaceseseo-gis-reieo-deiteoseseul-http-ringkeuro/korshunov.aiko2026-07-16T03:12:40.722508+00:00사용자가 Hugging Face Spaces에서 GIS 레이어 데이터셋을 HTTP 링크로 호스팅하는 방법을 묻습니다 https://korshunov.ai/zh-hans/article/12278-yong-hu-xun-wen-ru-he-zai-hugging-face-spaces-shang-jiang-gis-tu-ceng-shu-ju-ji/korshunov.aizh-cn2026-07-16T03:12:40.722508+00:00用户询问如何在 Hugging Face Spaces 上将 GIS 图层数据集作为 HTTP 链接托管 https://korshunov.ai/hi/article/12278-upyogkrtaa-puuchtaa-hai-ki-hugging-face-spaces-pr-gis-leyr-ddettaasett-ko-http/korshunov.aihi2026-07-16T03:12:40.722508+00:00उपयोगकर्ता पूछता है कि Hugging Face Spaces पर GIS लेयर डेटासेट को HTTP लिंक के रूप में कैसे होस्ट करें https://korshunov.ai/ja/article/12278-yu-sa-hahugging-face-spacestegisreiya-te-tawohttprinkutoshitehosutosurufang-fa/korshunov.aija2026-07-16T03:12:40.722508+00:00ユーザーはHugging Face SpacesでGISレイヤーデータをHTTPリンクとしてホストする方法を尋ねる https://korshunov.ai/id/article/12278-pengguna-bertanya-cara-menghosting-dataset-lapisan-gis-sebagai-tautan-http-di/korshunov.aiid2026-07-16T03:12:40.722508+00:00Pengguna bertanya cara menghosting dataset lapisan GIS sebagai tautan HTTP di Hugging Face Spaces https://korshunov.ai/ar/article/12278-ysl-lmstkhdm-n-kyfy-stdf-mjmw-t-bynt-tbqt-gis-krbt-http-l-msht-hugging-face/korshunov.aiar2026-07-16T03:12:40.722508+00:00يسأل المستخدم عن كيفية استضافة مجموعات بيانات طبقات GIS كرابط HTTP على مساحات Hugging Face https://korshunov.ai/en/article/12278-user-asks-how-to-host-gis-layer-datasets-as-http-links-on-hugging-face-spaces/korshunov.aien2026-07-16T03:12:40.722508+00:00User asks how to host GIS layer datasets as HTTP links on Hugging Face Spaces https://korshunov.ai/zh-hans/article/12236-relai-vcl-tong-guo-chi-xu-xue-xi-shi-zhi-neng-ti-you-hua-qi-zeng-yi-fu-he/korshunov.aizh-cn2026-07-16T03:07:56.128264+00:00RELAI-VCL 通过持续学习使智能体-优化器增益复合 https://korshunov.ai/id/article/12236-relai-vcl-menggandakan-keuntungan-agen-optimizer-melalui-pembelajaran/korshunov.aiid2026-07-16T03:07:56.128264+00:00RELAI-VCL menggandakan keuntungan agen-optimizer melalui pembelajaran berkelanjutan https://korshunov.ai/ar/article/12236-td-f-relai-vcl-mksb-thsyn-lwkyl-br-lt-lm-lmstmr/korshunov.aiar2026-07-16T03:07:56.128264+00:00تضاعف RELAI-VCL مكاسب تحسين الوكيل عبر التعلم المستمر https://korshunov.ai/fr/article/12236-relai-vcl-amplifie-les-gains-de-l-optimisation-d-agents-par-apprentissage/korshunov.aifr2026-07-16T03:07:56.128264+00:00RELAI-VCL amplifie les gains de l'optimisation d'agents par apprentissage continu https://korshunov.ai/ko/article/12236-relai-vcleun-jisogjeog-hagseubeul-tonghae-eijeonteu-coejeoghwa-ideugeul/korshunov.aiko2026-07-16T03:07:56.128264+00:00RELAI-VCL은 지속적 학습을 통해 에이전트 최적화 이득을 복합화합니다 https://korshunov.ai/ru/article/12236-relai-vcl-usilivaet-effekt-nakopleniia-preimushchestv-agenta-optimizatora-za/korshunov.airu2026-07-16T03:07:56.128264+00:00RELAI-VCL усиливает эффект накопления преимуществ агента-оптимизатора за счёт непрерывного обучения https://korshunov.ai/en/article/12236-relai-vcl-compounds-agent-optimizer-gains-via-continual-learning/korshunov.aien2026-07-16T03:07:56.128264+00:00RELAI-VCL compounds agent-optimizer gains via continual learning https://korshunov.ai/es/article/12236-relai-vcl-compuesta-las-ganancias-del-agente-optimizador-mediante-aprendizaje/korshunov.aies2026-07-16T03:07:56.128264+00:00RELAI-VCL compuesta las ganancias del agente-optimizador mediante aprendizaje continuo https://korshunov.ai/pt-br/article/12236-relai-vcl-compoe-ganhos-do-agente-otimizador-via-aprendizado-continuo/korshunov.aipt2026-07-16T03:07:56.128264+00:00RELAI-VCL compõe ganhos do agente-otimizador via aprendizado contínuo https://korshunov.ai/hi/article/12236-relai-vcl-nirntr-shikssnn-ke-maadhym-se-ejentt-onpttimaaijr-laabhon-ko-snyojit/korshunov.aihi2026-07-16T03:07:56.128264+00:00RELAI-VCL निरंतर शिक्षण के माध्यम से एजेंट-ऑप्टिमाइजर लाभों को संयोजित करता है https://korshunov.ai/ja/article/12236-relai-vclhaji-sok-de-xue-xi-niyorue-shientozui-shi-hua-noli-yi-wofu-he-saseru/korshunov.aija2026-07-16T03:07:56.128264+00:00RELAI-VCLは継続的学習によるエージェント最適化の利益を複合させる https://korshunov.ai/es/article/12235-replanteamiento-de-las-pruebas-de-penetracion-para-sistemas-de-ia-como-orientada/korshunov.aies2026-07-16T02:59:50.005865+00:00Replanteamiento de las pruebas de penetración para sistemas de IA como evaluación conductual orientada a objetivos https://korshunov.ai/ru/article/12235-pereosmyslenie-testirovaniia-na-proniknovenie-dlia-ii-sistem-kak-obektivno/korshunov.airu2026-07-16T02:59:50.005865+00:00Переосмысление тестирования на проникновение для ИИ-систем как объективно-ориентированной поведенческой оценки https://korshunov.ai/en/article/12235-reframing-penetration-testing-for-ai-systems-as-objective-driven-behavioral/korshunov.aien2026-07-16T02:59:50.005865+00:00Reframing penetration testing for AI systems as objective-driven behavioral evaluation https://korshunov.ai/zh-hans/article/12235-jiang-zhen-dui-aixi-tong-de-shen-tou-ce-shi-zhong-gou-wei-yi-mu-biao-wei-dao-de/korshunov.aizh-cn2026-07-16T02:59:50.005865+00:00将针对AI系统的渗透测试重构为以目标为导向的行为评估 https://korshunov.ai/pt-br/article/12235-reenquadramento-do-teste-de-penetracao-para-sistemas-de-ia-como-avaliacao-a/korshunov.aipt2026-07-16T02:59:50.005865+00:00Reenquadramento do teste de penetração para sistemas de IA como avaliação comportamental orientada a objetivos https://korshunov.ai/ko/article/12235-ai-siseuteme-daehan-cimtu-teseuteureul-mogjeog-giban-haengdong-pyeonggaro/korshunov.aiko2026-07-16T02:59:50.005865+00:00AI 시스템에 대한 침투 테스트를 목적 기반 행동 평가로 재정의 https://korshunov.ai/ja/article/12235-aishisutemunohenetore-shiyontesutowomu-de-qu-dong-xing-xing-dong-ping-jia-ding/korshunov.aija2026-07-16T02:59:50.005865+00:00AIシステムのペネトレーションテストを目的駆動型行動評価として再定義する https://korshunov.ai/hi/article/12235-ai-sisttm-ke-lie-pennettreshn-ttestting-ko-uddeshy-kendrit-vyvhaarik-muulyaankn/korshunov.aihi2026-07-16T02:59:50.005865+00:00AI सिस्टम के लिए पेंनेट्रेशन टेस्टिंग को उद्देश्य-केंद्रित व्यवहारिक मूल्यांकन के रूप में पुनर्परिभाषित करना https://korshunov.ai/id/article/12235-membingkai-ulang-pengujian-penetrasi-untuk-sistem-ai-sebagai-evaluasi-perilaku/korshunov.aiid2026-07-16T02:59:50.005865+00:00Membingkai ulang pengujian penetrasi untuk sistem AI sebagai evaluasi perilaku berbasis tujuan https://korshunov.ai/fr/article/12235-reformuler-les-tests-de-penetration-pour-les-systemes-d-ia-en-evaluation/korshunov.aifr2026-07-16T02:59:50.005865+00:00Reformuler les tests de pénétration pour les systèmes d'IA en évaluation comportementale orientée objectif https://korshunov.ai/ar/article/12235-d-syg-khtbr-lkhtrq-lnzm-ldhk-lstn-y-ktqyym-slwky-qym-l-lhdf/korshunov.aiar2026-07-16T02:59:50.005865+00:00إعادة صياغة اختبار الاختراق لأنظمة الذكاء الاصطناعي كتقييم سلوكي قائم على الأهداف https://korshunov.ai/ar/article/12234-tsmym-ktl-ltgdhy-lmmy-fy-lmhwwl-yhfz-l-rtb-ltdrj-br-l-mq/korshunov.aiar2026-07-16T02:52:25.923930+00:00تصميم كتلة التغذية الأمامية في المحوّل يحافظ على رتبة التدرج عبر العمق https://korshunov.ai/id/article/12234-desain-blok-feedforward-transformer-mempertahankan-rank-gradien-sepanjang/korshunov.aiid2026-07-16T02:52:25.923930+00:00Desain blok feedforward Transformer mempertahankan rank gradien sepanjang kedalaman https://korshunov.ai/fr/article/12234-la-conception-du-bloc-feedforward-transformer-preserve-le-rang-du-gradient-sur/korshunov.aifr2026-07-16T02:52:25.923930+00:00La conception du bloc feedforward Transformer préserve le rang du gradient sur la profondeur https://korshunov.ai/ko/article/12234-transformer-sunjeonpa-beulrog-seolgyeneun-gipie-geolcyeo-giulgi-raengkeureul/korshunov.aiko2026-07-16T02:52:25.923930+00:00Transformer 순전파 블록 설계는 깊이에 걸쳐 기울기 랭크를 보존한다 https://korshunov.ai/ja/article/12234-transformernoshun-chuan-bo-hurotsukushe-ji-ha-shen-saniwatatsutegou-pei-chi-suru/korshunov.aija2026-07-16T02:52:25.923930+00:00Transformerの順伝播ブロック設計は、深さにわたって勾配のランクを保持する https://korshunov.ai/pt-br/article/12234-o-design-do-bloco-feedforward-do-transformer-preserva-a-rank-do-gradiente-ao-da/korshunov.aipt2026-07-16T02:52:25.923930+00:00O design do bloco feedforward do Transformer preserva a rank do gradiente ao longo da profundidade https://korshunov.ai/es/article/12234-el-diseno-del-bloque-feedforward-del-transformer-preserva-el-rango-del-gradiente/korshunov.aies2026-07-16T02:52:25.923930+00:00El diseño del bloque feedforward del Transformer preserva el rango del gradiente a través de la profundidad https://korshunov.ai/ru/article/12234-dizain-bloka-priamoi-peredachi-transformera-sokhraniaet-rang-gradienta-na/korshunov.airu2026-07-16T02:52:25.923930+00:00Дизайн блока прямой передачи трансформера сохраняет ранг градиента на глубине https://korshunov.ai/en/article/12234-transformer-feedforward-block-design-preserves-gradient-rank-across-depth/korshunov.aien2026-07-16T02:52:25.923930+00:00Transformer feedforward block design preserves gradient rank across depth https://korshunov.ai/hi/article/12234-transformer-phiiddphonrvrdd-blonk-ddijaain-ghraaii-bhr-men-greddientt-raink-ko/korshunov.aihi2026-07-16T02:52:25.923930+00:00Transformer फ़ीडफॉरवर्ड ब्लॉक डिज़ाइन गहराई भर में ग्रेडिएंट रैंक को बनाए रखता है https://korshunov.ai/zh-hans/article/12234-transformerqian-kui-kuai-she-ji-zai-shen-du-shang-bao-chi-ti-du-zhi/korshunov.aizh-cn2026-07-16T02:52:25.923930+00:00Transformer前馈块设计在深度上保持梯度秩 https://korshunov.ai/es/article/12233-csfs-reduce-el-costo-computacional-en-un-21-en-la-prediccion-de-energia/korshunov.aies2026-07-16T02:43:21.812829+00:00CSFS reduce el costo computacional en un 21% en la predicción de energía renovable https://korshunov.ai/hi/article/12233-csfs-nviikrnniiy-uurjaa-puurvaanumaan-men-21-kii-gnnnaatmk-laagt-km-krtaa-hai/korshunov.aihi2026-07-16T02:43:21.812829+00:00CSFS नवीकरणीय ऊर्जा पूर्वानुमान में 21% की गणनात्मक लागत कम करता है https://korshunov.ai/ru/article/12233-csfs-snizhaet-vychislitelnye-zatraty-na-21-pri-prognozirovanii-vozobnovliaemoi/korshunov.airu2026-07-16T02:43:21.812829+00:00CSFS снижает вычислительные затраты на 21% при прогнозировании возобновляемой энергии https://korshunov.ai/ko/article/12233-csfs-jaesaengeneoji-yeceugeseo-gyesan-biyong-21-jeolgam/korshunov.aiko2026-07-16T02:43:21.812829+00:00CSFS, 재생에너지 예측에서 계산 비용 21% 절감 https://korshunov.ai/zh-hans/article/12233-csfszai-ke-zai-sheng-neng-yuan-yu-ce-zhong-jiang-ji-suan-cheng-ben-jiang-di-21/korshunov.aizh-cn2026-07-16T02:43:21.812829+00:00CSFS在可再生能源预测中将计算成本降低21% https://korshunov.ai/ja/article/12233-csfskazai-sheng-ke-neng-eneruki-yu-ce-nioiteji-suan-kosutowo21-xue-jian/korshunov.aija2026-07-16T02:43:21.812829+00:00CSFSが再生可能エネルギー予測において計算コストを21%削減 https://korshunov.ai/fr/article/12233-csfs-reduit-le-cout-computationnel-de-21-dans-la-prediction-des-energies/korshunov.aifr2026-07-16T02:43:21.812829+00:00CSFS réduit le coût computationnel de 21 % dans la prédiction des énergies renouvelables https://korshunov.ai/en/article/12233-csfs-reduces-computational-cost-by-21-in-renewable-energy-prediction/korshunov.aien2026-07-16T02:43:21.812829+00:00CSFS reduces computational cost by 21% in renewable energy prediction https://korshunov.ai/id/article/12233-csfs-mengurangi-biaya-komputasi-sebesar-21-dalam-prediksi-energi-terbarukan/korshunov.aiid2026-07-16T02:43:21.812829+00:00CSFS mengurangi biaya komputasi sebesar 21% dalam prediksi energi terbarukan https://korshunov.ai/ar/article/12233-csfs-yqll-ltklf-lhsby-bnsb-21-fy-twq-t-ltq-lmtjdd/korshunov.aiar2026-07-16T02:43:21.812829+00:00CSFS يقلل التكلفة الحسابية بنسبة 21% في توقعات الطاقة المتجددة https://korshunov.ai/pt-br/article/12233-csfs-reduz-o-custo-computacional-em-21-na-previsao-de-energia-renovavel/korshunov.aipt2026-07-16T02:43:21.812829+00:00CSFS reduz o custo computacional em 21% na previsão de energia renovável https://korshunov.ai/zh-hans/article/12232-yan-jiu-fa-xian-zhi-neng-ti-bian-ma-gong-ju-de-cai-yong-jin-xian-yu-githubxiang/korshunov.aizh-cn2026-07-16T02:35:54.522069+00:00研究发现，智能体编码工具的采用仅限于GitHub项目的一小部分 https://korshunov.ai/ja/article/12232-e-shientoxing-ko-teinkutsu-runocai-yong-hagithubhuroshiekutonokokuyi-bu-nixian/korshunov.aija2026-07-16T02:35:54.522069+00:00エージェント型コーディングツールの採用はGitHubプロジェクトのごく一部に限定されていることを研究が明らかに https://korshunov.ai/ar/article/12232-drs-tjd-n-tmd-dwt-lbrmj-lwkyly-mhdwd-bmjmw-fr-y-sgyr-mn-mshry-github/korshunov.aiar2026-07-16T02:35:54.522069+00:00دراسة تجد أن اعتماد أدوات البرمجة الوكيلية محدود بمجموعة فرعية صغيرة من مشاريع GitHub https://korshunov.ai/ko/article/12232-eijeonteu-koding-dogu-caetaegeun-github-peurojegteuyi-sosuegeman-jehandoemeul/korshunov.aiko2026-07-16T02:35:54.522069+00:00에이전트 코딩 도구 채택은 GitHub 프로젝트의 소수에게만 제한됨을 연구가 발견 https://korshunov.ai/fr/article/12232-une-etude-revele-que-l-adoption-des-outils-de-codage-agents-se-limite-a-un-petit/korshunov.aifr2026-07-16T02:35:54.522069+00:00Une étude révèle que l'adoption des outils de codage agents se limite à un petit sous-ensemble de projets GitHub https://korshunov.ai/es/article/12232-un-estudio-revela-que-la-adopcion-de-herramientas-de-codificacion-agentica-se-a/korshunov.aies2026-07-16T02:35:54.522069+00:00Un estudio revela que la adopción de herramientas de codificación agéntica se limita a un pequeño subconjunto de proyectos de GitHub https://korshunov.ai/id/article/12232-studi-menemukan-adopsi-alat-koding-agentic-terbatas-pada-subset-kecil-proyek/korshunov.aiid2026-07-16T02:35:54.522069+00:00Studi menemukan adopsi alat koding agentic terbatas pada subset kecil proyek GitHub https://korshunov.ai/pt-br/article/12232-estudo-descobre-que-a-adocao-de-ferramentas-de-codificacao-agentica-e-limitada-a/korshunov.aipt2026-07-16T02:35:54.522069+00:00Estudo descobre que a adoção de ferramentas de codificação agêntica é limitada a um pequeno subconjunto de projetos do GitHub https://korshunov.ai/hi/article/12232-adhyyn-paataa-hai-ki-ejenttik-kodding-ttuul-kaa-apnaanaa-github-priyojnaaon-ke/korshunov.aihi2026-07-16T02:35:54.522069+00:00अध्ययन पाता है कि एजेंटिक कोडिंग टूल का अपनाना GitHub परियोजनाओं के एक छोटे समूह तक सीमित है https://korshunov.ai/ru/article/12232-issledovanie-pokazyvaet-chto-vnedrenie-agentnykh-instrumentov-dlia-ogranicheno/korshunov.airu2026-07-16T02:35:54.522069+00:00Исследование показывает, что внедрение агентных инструментов для программирования ограничено небольшой частью проектов на GitHub https://korshunov.ai/en/article/12232-study-finds-agentic-coding-tool-adoption-is-limited-to-small-subset-of-github/korshunov.aien2026-07-16T02:35:54.522069+00:00Study finds agentic coding tool adoption is limited to small subset of GitHub projects https://korshunov.ai/ar/article/12231-twjyh-lkhbr-lmt-ddyn-yhsn-dq-lt-rf-ldwyy-l-lhrwf-lmnshwry-br-nmt-ktb-mmyz/korshunov.aiar2026-07-16T02:28:16.712001+00:00توجيه الخبراء المتعددين يحسن دقة التعرف الضوئي على الحروف المنشورية عبر أنماط كتابة مميزة https://korshunov.ai/id/article/12231-rute-multi-ahli-meningkatkan-akurasi-ocr-manchu-di-berbagai-gaya-tulisan-yang/korshunov.aiid2026-07-16T02:28:16.712001+00:00Rute Multi-ahli meningkatkan akurasi OCR Manchu di berbagai gaya tulisan yang berbeda https://korshunov.ai/ko/article/12231-dajung-jeonmunga-rautingeun-dayanghan-seoce-jeonbaneseo-manjueo-ocr/korshunov.aiko2026-07-16T02:28:16.712001+00:00다중 전문가 라우팅은 다양한 서체 전반에서 만주어 OCR 정확도를 향상시킵니다 https://korshunov.ai/ja/article/12231-maruchiekisuha-toru-teinkuha-yi-narushu-feng-quan-ti-teman-wen-ocrnojing-du/korshunov.aija2026-07-16T02:28:16.712001+00:00マルチエキスパートルーティングは、異なる書風全体で満文OCRの精度を向上させる https://korshunov.ai/hi/article/12231-bhu-vishessjny-raautting-vibhinn-lekhn-shailiyon-men-mncuu-ocr-sttiiktaa-ko-hai/korshunov.aihi2026-07-16T02:28:16.712001+00:00बहु-विशेषज्ञ राउटिंग विभिन्न लेखन शैलियों में मंचू OCR सटीकता को बढ़ाती है https://korshunov.ai/pt-br/article/12231-roteamento-de-multiplos-especialistas-melhora-a-precisidade-do-ocr-manchu-em-de/korshunov.aipt2026-07-16T02:28:16.712001+00:00Roteamento de Múltiplos Especialistas melhora a precisidade do OCR manchu em distintos estilos de escrita https://korshunov.ai/es/article/12231-el-enrutamiento-de-multiples-expertos-mejora-la-precision-del-ocr-manchu-en-de/korshunov.aies2026-07-16T02:28:16.712001+00:00El enrutamiento de múltiples expertos mejora la precisión del OCR manchu en distintos estilos de escritura https://korshunov.ai/ru/article/12231-marshrutizatsiia-s-uchastiem-neskolkikh-ekspertov-povyshaet-tochnost-teksta-ocr/korshunov.airu2026-07-16T02:28:16.712001+00:00Маршрутизация с участием нескольких экспертов повышает точность распознавания маньчжурского текста (OCR) при различных стилях письма https://korshunov.ai/en/article/12231-multi-expert-routing-improves-manchu-ocr-accuracy-across-distinct-writing-styles/korshunov.aien2026-07-16T02:28:16.712001+00:00Multi-Expert Routing improves Manchu OCR accuracy across distinct writing styles https://korshunov.ai/zh-hans/article/12231-duo-zhuan-jia-lu-you-ti-gao-liao-man-wen-ocrzai-bu-tong-shu-xie-feng-ge-xia-de/korshunov.aizh-cn2026-07-16T02:28:16.712001+00:00多专家路由提高了满文OCR在不同书写风格下的准确率 https://korshunov.ai/fr/article/12231-le-routage-multi-experts-ameliore-la-precision-de-l-ocr-mandchou-sur-des-styles/korshunov.aifr2026-07-16T02:28:16.712001+00:00Le routage multi-experts améliore la précision de l'OCR mandchou sur des styles d'écriture distincts https://korshunov.ai/id/article/12230-kerangka-kerja-yang-dipercepat-ai-memvalidasi-peningkatan-keterampilan-cepat-di/korshunov.aiid2026-07-16T02:19:59.006759+00:00Kerangka kerja yang dipercepat AI memvalidasi peningkatan keterampilan profesional cepat di lima tahap https://korshunov.ai/ja/article/12230-aijia-su-hure-muwa-kuka5tsunoduan-jie-niwataruxun-su-nazhuan-men-de-sukiruxiang/korshunov.aija2026-07-16T02:19:59.006759+00:00AI加速フレームワークが5つの段階にわたる迅速な専門的スキル向上を検証 https://korshunov.ai/zh-hans/article/12230-aijia-su-kuang-jia-zai-wu-ge-jie-duan-yan-zheng-liao-kuai-su-de-zhuan-ye-ji-neng/korshunov.aizh-cn2026-07-16T02:19:59.006759+00:00AI加速框架在五个阶段验证了快速的专业技能提升 https://korshunov.ai/pt-br/article/12230-framework-acelerado-por-ia-valida-rapida-atualizacao-profissional-em-cinco/korshunov.aipt2026-07-16T02:19:59.006759+00:00Framework acelerado por IA valida rápida atualização profissional em cinco etapas https://korshunov.ai/ko/article/12230-ai-gasog-peureimweokeuga-5dangyee-geolcin-sinsoghan-jeonmun-gisul-hyangsang/korshunov.aiko2026-07-16T02:19:59.006759+00:00AI 가속 프레임워크가 5단계에 걸친 신속한 전문 기술 향상 검증 https://korshunov.ai/hi/article/12230-ai-tvrit-phremvrk-paanc-crnnon-men-tvrit-peshevr-kaushl-vrddhi-kii-styaapn-krtaa/korshunov.aihi2026-07-16T02:19:59.006759+00:00AI-त्वरित फ्रेमवर्क पांच चरणों में त्वरित पेशेवर कौशल वृद्धि की सत्यापन करता है https://korshunov.ai/en/article/12230-ai-accelerated-framework-validates-rapid-professional-upskilling-across-five/korshunov.aien2026-07-16T02:19:59.006759+00:00AI-accelerated framework validates rapid professional upskilling across five stages https://korshunov.ai/es/article/12230-un-marco-acelerado-por-ia-valida-la-rapida-actualizacion-profesional-en-cinco/korshunov.aies2026-07-16T02:19:59.006759+00:00Un marco acelerado por IA valida la rápida actualización profesional en cinco etapas https://korshunov.ai/fr/article/12230-un-cadre-accelere-par-l-ia-valide-la-montee-en-competences-professionnelles-sur/korshunov.aifr2026-07-16T02:19:59.006759+00:00Un cadre accéléré par l'IA valide la montée en compétences professionnelles rapide sur cinq étapes https://korshunov.ai/ru/article/12230-freimvork-s-ii-uskoreniem-podtverzhdaet-bystroe-povyshenie-kvalifikatsii-na/korshunov.airu2026-07-16T02:19:59.006759+00:00Фреймворк с ИИ-ускорением подтверждает быстрое повышение квалификации специалистов на пяти этапах https://korshunov.ai/ar/article/12230-tr-m-zz-bldhk-lstn-y-ythbt-tsry-ttwyr-lmhrt-lmhny-br-khms-mrhl/korshunov.aiar2026-07-16T02:19:59.006759+00:00إطار معزز بالذكاء الاصطناعي يثبت تسريع تطوير المهارات المهنية عبر خمس مراحل https://korshunov.ai/pt-br/article/12229-earthquaker-ai-framework-rag-com-avaliacao-baseada-em-rubricas-para-educacao/korshunov.aipt2026-07-16T02:12:36.505298+00:00Earthquaker-AI: framework RAG com avaliação baseada em rubricas para educação primária sobre terremotos https://korshunov.ai/zh-hans/article/12229-earthquaker-ai-yong-yu-xiao-xue-di-zhen-jiao-yu-de-ji-yu-liang-gui-ping-gu-de/korshunov.aizh-cn2026-07-16T02:12:36.505298+00:00Earthquaker-AI：用于小学地震教育的基于量规评估的RAG框架 https://korshunov.ai/ru/article/12229-earthquaker-ai-sistema-rag-s-otsenkoi-po-rubrikam-dlia-obucheniia-detei-shkoly/korshunov.airu2026-07-16T02:12:36.505298+00:00Earthquaker-AI: система RAG с оценкой по рубрикам для обучения детей начальной школы землетрясениям https://korshunov.ai/es/article/12229-earthquaker-ai-marco-rag-con-evaluacion-basada-en-rubricas-para-la-educacion/korshunov.aies2026-07-16T02:12:36.505298+00:00Earthquaker-AI: marco RAG con evaluación basada en rúbricas para la educación primaria sobre terremotos https://korshunov.ai/en/article/12229-earthquaker-ai-rag-framework-with-rubric-based-assessment-for-primary-school/korshunov.aien2026-07-16T02:12:36.505298+00:00Earthquaker-AI: RAG framework with rubric-based assessment for primary school earthquake education https://korshunov.ai/hi/article/12229-earthquaker-ai-praathmik-vidyaaly-bhuuknp-shikssaa-ke-lie-ruubrik-aadhaarit-ke/korshunov.aihi2026-07-16T02:12:36.505298+00:00Earthquaker-AI: प्राथमिक विद्यालय भूकंप शिक्षा के लिए रूब्रिक-आधारित मूल्यांकन के साथ RAG फ्रेमवर्क https://korshunov.ai/ja/article/12229-earthquaker-ai-xiao-xue-xiao-nodi-zhen-jiao-yu-notamenoji-zhun-he-suping-jia-fu/korshunov.aija2026-07-16T02:12:36.505298+00:00Earthquaker-AI: 小学校の地震教育のための基準ベース評価付きRAGフレームワーク https://korshunov.ai/ar/article/12229-earthquaker-ai-tr-ml-rag-m-tqyym-qym-l-lm-yyr-lt-lym-lzlzl-fy-lmrhl-lbtdyy/korshunov.aiar2026-07-16T02:12:36.505298+00:00Earthquaker-AI: إطار عمل RAG مع تقييم قائم على المعايير لتعليم الزلازل في المرحلة الابتدائية https://korshunov.ai/ko/article/12229-earthquaker-ai-codeunghaggyo-jijin-gyoyugeul-wihan-gijun-giban-pyeongga-rag/korshunov.aiko2026-07-16T02:12:36.505298+00:00Earthquaker-AI: 초등학교 지진 교육을 위한 기준 기반 평가 RAG 프레임워크 https://korshunov.ai/fr/article/12229-earthquaker-ai-cadre-rag-avec-evaluation-basee-sur-des-criteres-pour-l-education/korshunov.aifr2026-07-16T02:12:36.505298+00:00Earthquaker-AI : cadre RAG avec évaluation basée sur des critères pour l'éducation sismique primaire https://korshunov.ai/id/article/12229-earthquaker-ai-kerangka-kerja-rag-dengan-penilaian-berbasis-rubrik-untuk-gempa/korshunov.aiid2026-07-16T02:12:36.505298+00:00Earthquaker-AI: Kerangka kerja RAG dengan penilaian berbasis rubrik untuk pendidikan gempa bumi sekolah dasar https://korshunov.ai/ko/article/12228-deep-interactioneul-tonghae-llm-curon-oryureul-jeongmilhage-ingani-sujeong/korshunov.aiko2026-07-16T02:02:35.451910+00:00Deep Interaction을 통해 LLM 추론 오류를 정밀하게 인간이 수정 가능 https://korshunov.ai/fr/article/12228-deep-interaction-permet-une-correction-humaine-precise-des-erreurs-de-des-llm/korshunov.aifr2026-07-16T02:02:35.451910+00:00Deep Interaction permet une correction humaine précise des erreurs de raisonnement des LLM https://korshunov.ai/id/article/12228-deep-interaction-memungkinkan-koreksi-manusia-yang-presisi-atas-kesalahan-llm/korshunov.aiid2026-07-16T02:02:35.451910+00:00Deep Interaction memungkinkan koreksi manusia yang presisi atas kesalahan penalaran LLM https://korshunov.ai/ar/article/12228-ytyh-deep-interaction-tshyhan-bshryan-dqyqan-lkht-lstdll-fy-nmdhj-llgt-lkbyr/korshunov.aiar2026-07-16T02:02:35.451910+00:00يتيح Deep Interaction تصحيحًا بشريًا دقيقًا لأخطاء الاستدلال في نماذج اللغات الكبيرة https://korshunov.ai/pt-br/article/12228-interacao-profunda-permite-correcao-precisa-de-erros-de-raciocinio-de-llm-por/korshunov.aipt2026-07-16T02:02:35.451910+00:00Interação profunda permite correção precisa de erros de raciocínio de LLM por humanos https://korshunov.ai/hi/article/12228-ghraa-inttraikshn-llm-trk-truttiyon-kaa-sttiik-maanv-sudhaar-skssm-bnaataa-hai/korshunov.aihi2026-07-16T02:02:35.451910+00:00गहरा इंटरैक्शन LLM तर्क त्रुटियों का सटीक मानव सुधार सक्षम बनाता है https://korshunov.ai/es/article/12228-la-interaccion-profunda-permite-la-correccion-precisa-de-errores-de-razonamiento/korshunov.aies2026-07-16T02:02:35.451910+00:00La interacción profunda permite la corrección precisa de errores de razonamiento de LLM por humanos https://korshunov.ai/ja/article/12228-deep-interactionniyorillmnotui-lun-era-wozheng-que-niren-jian-kaxiu-zheng-ke-ni/korshunov.aija2026-07-16T02:02:35.451910+00:00Deep InteractionによりLLMの推論エラーを正確に人間が修正可能に https://korshunov.ai/zh-hans/article/12228-shen-du-jiao-hu-shi-xian-ren-lei-dui-llmtui-li-cuo-wu-de-jing-que-xiu-zheng/korshunov.aizh-cn2026-07-16T02:02:35.451910+00:00深度交互实现人类对LLM推理错误的精确修正 https://korshunov.ai/en/article/12228-deep-interaction-enables-precise-human-correction-of-llm-reasoning-errors/korshunov.aien2026-07-16T02:02:35.451910+00:00Deep Interaction enables precise human correction of LLM reasoning errors https://korshunov.ai/ru/article/12228-glubokoe-vzaimodeistvie-obespechivaet-tochnoe-ispravlenie-oshibok-rassuzhdeniia/korshunov.airu2026-07-16T02:02:35.451910+00:00Глубокое взаимодействие обеспечивает точное исправление ошибок рассуждения LLM человеком https://korshunov.ai/zh-hans/article/12227-simon-willison-jiang-grok-cli-de-mermaid-xuan-ran-qi-bian-yi-wei-webassembly/korshunov.aizh-cn2026-07-16T01:56:21.855728+00:00Simon Willison 将 Grok CLI 的 Mermaid 渲染器编译为 WebAssembly https://korshunov.ai/en/article/12227-simon-willison-compiles-grok-cli-s-mermaid-renderer-to-webassembly/korshunov.aien2026-07-16T01:56:21.855728+00:00Simon Willison compiles Grok CLI's Mermaid renderer to WebAssembly https://korshunov.ai/ar/article/12227-symwn-wylyswn-yhwwl-murasm-mermaid-lkhs-b-grok-cli-l-webassembly/korshunov.aiar2026-07-16T01:56:21.855728+00:00سيمون ويليسون يحوّل مُرَسم Mermaid الخاص بـ Grok CLI إلى WebAssembly https://korshunov.ai/ru/article/12227-saimon-uillison-kompiliruet-renderer-mermaid-iz-grok-cli-v-webassembly/korshunov.airu2026-07-16T01:56:21.855728+00:00Саймон Уиллисон компилирует рендерер Mermaid из Grok CLI в WebAssembly https://korshunov.ai/ja/article/12227-simon-willisonkagrok-clinomermaidrentara-wowebassemblynikonhairu/korshunov.aija2026-07-16T01:56:21.855728+00:00Simon WillisonがGrok CLIのMermaidレンダラーをWebAssemblyにコンパイル https://korshunov.ai/es/article/12227-simon-willison-compila-el-renderizador-mermaid-de-grok-cli-a-webassembly/korshunov.aies2026-07-16T01:56:21.855728+00:00Simon Willison compila el renderizador Mermaid de Grok CLI a WebAssembly https://korshunov.ai/ko/article/12227-simon-willisoni-grok-cliyi-mermaid-rendeoreoreul-webassemblyro-keompail/korshunov.aiko2026-07-16T01:56:21.855728+00:00Simon Willison이 Grok CLI의 Mermaid 렌더러를 WebAssembly로 컴파일 https://korshunov.ai/hi/article/12227-saaimn-vilisn-ne-grok-cli-ke-mermaid-renddrr-ko-webassembly-men-knpaail-kiyaa/korshunov.aihi2026-07-16T01:56:21.855728+00:00साइमन विलिसन ने Grok CLI के Mermaid रेंडरर को WebAssembly में कंपाइल किया https://korshunov.ai/pt-br/article/12227-simon-willison-compila-o-renderizador-mermaid-do-grok-cli-para-webassembly/korshunov.aipt2026-07-16T01:56:21.855728+00:00Simon Willison compila o renderizador Mermaid do Grok CLI para WebAssembly https://korshunov.ai/fr/article/12227-simon-willison-compile-le-rendu-mermaid-de-grok-cli-en-webassembly/korshunov.aifr2026-07-16T01:56:21.855728+00:00Simon Willison compile le rendu Mermaid de Grok CLI en WebAssembly https://korshunov.ai/id/article/12227-simon-willison-mengompilasi-renderer-mermaid-dari-grok-cli-ke-webassembly/korshunov.aiid2026-07-16T01:56:21.855728+00:00Simon Willison mengompilasi renderer Mermaid dari Grok CLI ke WebAssembly https://korshunov.ai/ko/article/12093-vihorec-pumjil-gwanridoen-beteunameo-hotel-cuceon-deiteoses-mic-koldeu-seutateu/korshunov.aiko2026-07-16T01:49:22.274860+00:00ViHoRec: 품질 관리된 베트남어 호텔 추천 데이터셋 및 콜드 스타트 벤치마크 https://korshunov.ai/id/article/12093-vihorec-dataset-rekomendasi-hotel-vietnam-berkontrol-kualitas-dan-benchmark-cold/korshunov.aiid2026-07-16T01:49:22.274860+00:00ViHoRec: Dataset Rekomendasi Hotel Vietnam Berkontrol Kualitas dan Benchmark Cold-Start https://korshunov.ai/ar/article/12093-vihorec-mjmw-bynt-ltwsy-lfndq-lfytnmy-khd-llthkm-fy-ljwd-wm-yr-lbd-ltshgyl-lbrd/korshunov.aiar2026-07-16T01:49:22.274860+00:00ViHoRec: مجموعة بيانات لتوصية الفنادق الفيتنامية خاضعة للتحكم في الجودة ومعيار لبدء التشغيل البارد https://korshunov.ai/fr/article/12093-vihorec-un-jeu-de-donnees-de-recommandation-d-hotels-vietnamiens-controle-en-et/korshunov.aifr2026-07-16T01:49:22.274860+00:00ViHoRec : Un jeu de données de recommandation d'hôtels vietnamiens contrôlé en qualité et un benchmark de démarrage à froid https://korshunov.ai/pt-br/article/12093-vihorec-um-conjunto-de-dados-controlado-para-recomendacao-de-hoteis-no-vietna-e/korshunov.aipt2026-07-16T01:49:22.274860+00:00ViHoRec: Um Conjunto de Dados Controlado para Recomendação de Hotéis no Vietnã e Benchmark de Cold-Start https://korshunov.ai/hi/article/12093-vihorec-gunnvttaa-niyntrit-viytnaamii-hottl-anushnkaa-ddettaasett-aur-koldd/korshunov.aihi2026-07-16T01:49:22.274860+00:00ViHoRec: गुणवत्ता-नियंत्रित वियतनामी होटल अनुशंका डेटासेट और कोल्ड-स्टार्ट बेंचमार्क https://korshunov.ai/es/article/12093-vihorec-un-conjunto-de-datos-controlado-para-la-recomendacion-de-hoteles-en-y-de/korshunov.aies2026-07-16T01:49:22.274860+00:00ViHoRec: Un conjunto de datos controlado para la recomendación de hoteles en Vietnam y una prueba de referencia para el arranque en frío https://korshunov.ai/zh-hans/article/12093-vihorec-yi-ge-zhi-liang-ke-kong-de-yue-nan-jiu-dian-yu-ding-shu-ju-ji-ji-leng-qi/korshunov.aizh-cn2026-07-16T01:49:22.274860+00:00ViHoRec：一个质量可控的越南酒店预订数据集及冷启动基准 https://korshunov.ai/en/article/12093-vihorec-a-quality-controlled-vietnamese-hotel-recommendation-dataset-and-cold/korshunov.aien2026-07-16T01:49:22.274860+00:00ViHoRec: A Quality-Controlled Vietnamese Hotel Recommendation Dataset and Cold-Start Benchmark https://korshunov.ai/ru/article/12093-vihorec-kontroliruemyi-nabor-dannykh-dlia-rekomendatsii-otelei-vo-vetname-i-dlia/korshunov.airu2026-07-16T01:49:22.274860+00:00ViHoRec: Контролируемый набор данных для рекомендации отелей во Вьетнаме и бенчмарк для задачи холодного старта https://korshunov.ai/ja/article/12093-vihorec-pin-zhi-guan-li-saretahetonamuyu-hoterutui-jian-te-tasetsutotoko-ku/korshunov.aija2026-07-16T01:49:22.274860+00:00ViHoRec: 品質管理されたベトナム語ホテル推薦データセットとコールドスタートベンチマーク https://korshunov.ai/zh-hans/article/12141-shu-zhi-tian-qi-yu-bao-wu-chai-xia-shen-du-xue-xi-guang-fu-yu-ce-lu-bang-xing-de/korshunov.aizh-cn2026-07-16T01:40:49.926655+00:00数值天气预报误差下深度学习光伏预测鲁棒性的时空分析 https://korshunov.ai/pt-br/article/12141-analise-espaciotemporal-da-robustez-do-aprendizado-profundo-na-previsao-sob-de/korshunov.aipt2026-07-16T01:40:49.926655+00:00Análise espaciotemporal da robustez do aprendizado profundo na previsão fotovoltaica sob erros de previsão numérica do tempo https://korshunov.ai/ko/article/12141-nwp-oca-haeseo-dibreoning-pv-yeceugyi-ganggeonseonge-daehan-sigonggan-bunseog/korshunov.aiko2026-07-16T01:40:49.926655+00:00NWP 오차 하에서 딥러닝 PV 예측의 강건성에 대한 시공간 분석 https://korshunov.ai/ja/article/12141-nwpwu-chai-xia-niokerutei-hura-ninkupvyu-ce-noqiang-ren-xing-niguan-surushi-kong/korshunov.aija2026-07-16T01:40:49.926655+00:00NWP誤差下におけるディープラーニングPV予測の強靭性に関する時空間解析 https://korshunov.ai/hi/article/12141-nwp-truttiyon-ke-tht-ghn-shikssnn-pv-puurvaanumaan-kii-mjbuutii-kaa-sthaan-kaal/korshunov.aihi2026-07-16T01:40:49.926655+00:00NWP त्रुटियों के तहत गहन शिक्षण PV पूर्वानुमान की मजबूती का स्थान-काल विश्लेषण https://korshunov.ai/fr/article/12141-analyse-spatio-temporelle-de-la-robustesse-des-previsions-pv-par-apprentissage/korshunov.aifr2026-07-16T01:40:49.926655+00:00Analyse spatio-temporelle de la robustesse des prévisions PV par apprentissage profond sous erreurs NWP https://korshunov.ai/es/article/12141-analisis-espaciotemporal-de-la-robustez-del-aprendizaje-profundo-en-la-bajo-de/korshunov.aies2026-07-16T01:40:49.926655+00:00Análisis espaciotemporal de la robustez del aprendizaje profundo en la predicción fotovoltaica bajo errores de predicción meteorológica numérica https://korshunov.ai/id/article/12141-analisis-spasio-temporal-ketahanan-prakiraan-pv-pembelajaran-mendalam-di-bawah/korshunov.aiid2026-07-16T01:40:49.926655+00:00Analisis spasio-temporal ketahanan prakiraan PV pembelajaran mendalam di bawah kesalahan NWP https://korshunov.ai/ar/article/12141-thlyl-mkny-zmny-lmrwn-ltnbw-bltq-lkhrwdwyy-bstkhdm-lt-lm-l-myq-tht-kht-ltnbw-l/korshunov.aiar2026-07-16T01:40:49.926655+00:00تحليل مكاني زمني لمرونة التنبؤ بالطاقة الكهروضوئية باستخدام التعلم العميق تحت أخطاء التنبؤ العددي للطقس https://korshunov.ai/en/article/12141-spatiotemporal-analysis-of-deep-learning-pv-forecasting-robustness-under-nwp/korshunov.aien2026-07-16T01:40:49.926655+00:00Spatiotemporal analysis of deep learning PV forecasting robustness under NWP errors https://korshunov.ai/ru/article/12141-prostranstvenno-vremennoi-analiz-ustoichivosti-glubokogo-obucheniia-v-energii/korshunov.airu2026-07-16T01:40:49.926655+00:00Пространственно-временной анализ устойчивости глубокого обучения в прогнозировании фотоэлектрической энергии при ошибках численного прогнозирования погоды https://korshunov.ai/fr/article/12092-l-etude-pope-ne-trouve-aucune-preuve-que-le-contenu-d-erreur-aide-les-petits-de/korshunov.aifr2026-07-16T01:25:03.153185+00:00L'étude PoPE ne trouve aucune preuve que le contenu d'erreur aide les petits modèles de code gelés à s'auto-réparer https://korshunov.ai/id/article/12092-studi-pope-tidak-menemukan-bukti-bahwa-konten-kesalahan-membantu-model-kode-beku/korshunov.aiid2026-07-16T01:25:03.153185+00:00Studi PoPE tidak menemukan bukti bahwa konten kesalahan membantu model kode kecil beku memperbaiki diri https://korshunov.ai/ja/article/12092-popeyan-jiu-era-nei-rong-kadong-jie-saretaxiao-gui-mo-ko-tomoterunozi-ji-xiu-fu/korshunov.aija2026-07-16T01:25:03.153185+00:00PoPE研究：エラー内容が凍結された小規模コードモデルの自己修復に役立つ証拠は見つからなかった https://korshunov.ai/ar/article/12092-drs-pope-l-tjd-dlylan-l-n-mhtw-lkht-ys-d-nmdhj-lshfr-lsgyr-lmtjmd-l-lslh-ldhty/korshunov.aiar2026-07-16T01:25:03.153185+00:00دراسة PoPE لا تجد دليلاً على أن محتوى الخطأ يساعد نماذج الشفرة الصغيرة المتجمدة على الإصلاح الذاتي https://korshunov.ai/hi/article/12092-pope-adhyyn-men-koii-sbuut-nhiin-milaa-ki-trutti-saamgrii-jmii-huii-chotte-kodd/korshunov.aihi2026-07-16T01:25:03.153185+00:00PoPE अध्ययन में कोई सबूत नहीं मिला कि त्रुटि सामग्री जमी हुई छोटे कोड मॉडलों को स्वयं ठीक करने में मदद करती है https://korshunov.ai/pt-br/article/12092-estudo-pope-nao-encontra-evidencias-de-que-o-conteudo-do-erro-ajude-modelos-de-a/korshunov.aipt2026-07-16T01:25:03.153185+00:00Estudo PoPE não encontra evidências de que o conteúdo do erro ajude modelos pequenos de código congelados a se auto-repararem https://korshunov.ai/zh-hans/article/12092-popeyan-jiu-fa-xian-cuo-wu-nei-rong-wu-zhu-yu-dong-jie-de-xiao-xing-dai-ma-mo-zi/korshunov.aizh-cn2026-07-16T01:25:03.153185+00:00PoPE研究发现，错误内容无助于冻结的小型代码模型进行自我修复 https://korshunov.ai/ko/article/12092-pope-yeongu-oryu-naeyongi-donggyeoldoen-sohyeong-kodeu-modelyi-jagi-surie-doumi/korshunov.aiko2026-07-16T01:25:03.153185+00:00PoPE 연구: 오류 내용이 동결된 소형 코드 모델의 자기 수리에 도움이 된다는 증거 없음 https://korshunov.ai/en/article/12092-pope-study-finds-no-evidence-that-error-content-helps-frozen-small-code-models/korshunov.aien2026-07-16T01:25:03.153185+00:00PoPE study finds no evidence that error content helps frozen small code models self-repair https://korshunov.ai/ru/article/12092-issledovanie-pope-ne-obnaruzhilo-dokazatelstv-togo-chto-soderzhanie-oshibok-koda/korshunov.airu2026-07-16T01:25:03.153185+00:00Исследование PoPE не обнаружило доказательств того, что содержание ошибок помогает замороженным малым моделям кода самоисправляться https://korshunov.ai/es/article/12092-el-estudio-pope-no-encuentra-evidencia-de-que-el-contenido-del-error-ayude-a-los/korshunov.aies2026-07-16T01:25:03.153185+00:00El estudio PoPE no encuentra evidencia de que el contenido del error ayude a los modelos pequeños de código congelados a repararse solos https://korshunov.ai/fr/article/12222-nvidia-introduit-des-agents-ia-pour-developper-des-runtimes-usd-legers/korshunov.aifr2026-07-16T01:18:14.490989+00:00NVIDIA introduit des agents IA pour développer des runtimes USD légers https://korshunov.ai/ru/article/12222-nvidia-predstavliaet-ai-agentov-dlia-razrabotki-legkikh-ispolniaiushchikh-sred/korshunov.airu2026-07-16T01:18:14.490989+00:00NVIDIA представляет AI-агентов для разработки легких исполняющих сред USD https://korshunov.ai/es/article/12222-nvidia-presenta-agentes-de-ia-para-desarrollar-entornos-de-ejecucion-usd-ligeros/korshunov.aies2026-07-16T01:18:14.490989+00:00NVIDIA presenta agentes de IA para desarrollar entornos de ejecución USD ligeros https://korshunov.ai/en/article/12222-nvidia-introduces-ai-agents-to-develop-lightweight-usd-runtimes/korshunov.aien2026-07-16T01:18:14.490989+00:00NVIDIA introduces AI agents to develop lightweight USD runtimes https://korshunov.ai/pt-br/article/12222-nvidia-apresenta-agentes-de-ia-para-desenvolver-runtimes-usd-leves/korshunov.aipt2026-07-16T01:18:14.490989+00:00NVIDIA apresenta agentes de IA para desenvolver runtimes USD leves https://korshunov.ai/zh-hans/article/12222-nvidia-tui-chu-ai-dai-li-yi-kai-fa-qing-liang-ji-usd-yun-xing-shi/korshunov.aizh-cn2026-07-16T01:18:14.490989+00:00NVIDIA 推出 AI 代理以开发轻量级 USD 运行时 https://korshunov.ai/hi/article/12222-nvidia-ne-hlke-usd-rnttaaims-viksit-krne-ke-lie-ai-ejentts-kaa-pricy-diyaa/korshunov.aihi2026-07-16T01:18:14.490989+00:00NVIDIA ने हल्के USD रनटाइम्स विकसित करने के लिए AI एजेंट्स का परिचय दिया https://korshunov.ai/ko/article/12222-nvidia-gyeongryang-usd-reontaim-gaebaleul-wihan-ai-eijeonteu-doib/korshunov.aiko2026-07-16T01:18:14.490989+00:00NVIDIA, 경량 USD 런타임 개발을 위한 AI 에이전트 도입 https://korshunov.ai/ja/article/12222-nvidiakaaie-shientowoyong-itaqing-liang-usdrantaimunokai-fa-wodao-ru/korshunov.aija2026-07-16T01:18:14.490989+00:00NVIDIAがAIエージェントを用いた軽量USDランタイムの開発を導入 https://korshunov.ai/id/article/12222-nvidia-memperkenalkan-agen-ai-untuk-mengembangkan-runtime-usd-ringan/korshunov.aiid2026-07-16T01:18:14.490989+00:00NVIDIA memperkenalkan agen AI untuk mengembangkan runtime USD ringan https://korshunov.ai/ar/article/12222-nfydy-tqdm-wkl-ldhk-lstn-y-lttwyr-wqt-tshgyl-usd-khfyf/korshunov.aiar2026-07-16T01:18:14.490989+00:00إنفيديا تقدم وكلاء الذكاء الاصطناعي لتطوير وقت تشغيل USD خفيف https://korshunov.ai/id/article/12221-nvidia-deepstream-9-1-memungkinkan-pelacakan-3d-multi-kamera/korshunov.aiid2026-07-16T01:11:11.257613+00:00NVIDIA DeepStream 9.1 memungkinkan pelacakan 3D multi-kamera https://korshunov.ai/ar/article/12221-nvidia-deepstream-9-1-ytyh-lttb-thlthy-lb-d-mt-dd-lkmyrt/korshunov.aiar2026-07-16T01:11:11.257613+00:00NVIDIA DeepStream 9.1 يتيح التتبع ثلاثي الأبعاد متعدد الكاميرات https://korshunov.ai/fr/article/12221-nvidia-deepstream-9-1-permet-le-suivi-3d-multi-cameras/korshunov.aifr2026-07-16T01:11:11.257613+00:00NVIDIA DeepStream 9.1 permet le suivi 3D multi-caméras https://korshunov.ai/ko/article/12221-nvidia-deepstream-9-1-dajung-kamera-3d-cujeog-jiweon/korshunov.aiko2026-07-16T01:11:11.257613+00:00NVIDIA DeepStream 9.1, 다중 카메라 3D 추적 지원 https://korshunov.ai/ja/article/12221-nvidia-deepstream-9-1kamaruchikamera3dtoratsukinkuwoshi-xian/korshunov.aija2026-07-16T01:11:11.257613+00:00NVIDIA DeepStream 9.1がマルチカメラ3Dトラッキングを実現 https://korshunov.ai/hi/article/12221-nvidia-deepstream-9-1-bhu-kaimraa-3d-ttraiking-skssm-krtaa-hai/korshunov.aihi2026-07-16T01:11:11.257613+00:00NVIDIA DeepStream 9.1 बहु-कैमरा 3D ट्रैकिंग सक्षम करता है https://korshunov.ai/pt-br/article/12221-nvidia-deepstream-9-1-permite-rastreamento-3d-multicanal/korshunov.aipt2026-07-16T01:11:11.257613+00:00NVIDIA DeepStream 9.1 permite rastreamento 3D multicanal https://korshunov.ai/zh-hans/article/12221-nvidia-deepstream-9-1-zhi-chi-duo-she-xiang-tou-3d-zhui-zong/korshunov.aizh-cn2026-07-16T01:11:11.257613+00:00NVIDIA DeepStream 9.1 支持多摄像头 3D 追踪 https://korshunov.ai/en/article/12221-nvidia-deepstream-9-1-enables-multi-camera-3d-tracking/korshunov.aien2026-07-16T01:11:11.257613+00:00NVIDIA DeepStream 9.1 enables multi-camera 3D tracking https://korshunov.ai/ru/article/12221-nvidia-deepstream-9-1-obespechivaet-mnogokamernoe-3d-otslezhivanie/korshunov.airu2026-07-16T01:11:11.257613+00:00NVIDIA DeepStream 9.1 обеспечивает многокамерное 3D-отслеживание https://korshunov.ai/es/article/12221-nvidia-deepstream-9-1-habilita-el-seguimiento-3d-multicanal/korshunov.aies2026-07-16T01:11:11.257613+00:00NVIDIA DeepStream 9.1 habilita el seguimiento 3D multicanal https://korshunov.ai/zh-hans/article/12223-yong-hu-ce-shi-xian-shi-1-bit-hy3-liang-hua-zai-89-gb-xia-bao-liu-zhi-liang/korshunov.aizh-cn2026-07-16T01:04:00.101002+00:00用户测试显示 1-bit Hy3 量化在 89 GB 下保留质量 https://korshunov.ai/pt-br/article/12223-testes-de-usuarios-mostram-que-a-quantizacao-hy3-de-1-bit-mantem-a-qualidade-em/korshunov.aipt2026-07-16T01:04:00.101002+00:00Testes de usuários mostram que a quantização Hy3 de 1 bit mantém a qualidade em 89 GB https://korshunov.ai/es/article/12223-las-pruebas-de-usuarios-muestran-que-la-cuantizacion-hy3-de-1-bit-conserva-la-en/korshunov.aies2026-07-16T01:04:00.101002+00:00Las pruebas de usuarios muestran que la cuantización Hy3 de 1 bit conserva la calidad en 89 GB https://korshunov.ai/hi/article/12223-yuujr-ttestt-89-gb-pr-gunnvttaa-bnaae-rkhte-hue-1-bitt-hy3-kvaanttiikrnn/korshunov.aihi2026-07-16T01:04:00.101002+00:00यूज़र टेस्ट: 89 GB पर गुणवत्ता बनाए रखते हुए 1-बिट Hy3 क्वांटीकरण https://korshunov.ai/ja/article/12223-89-gbtepin-zhi-wowei-chi-suru1hitsutohy3liang-zi-hua-noyu-sa-tesuto/korshunov.aija2026-07-16T01:04:00.101002+00:0089 GBで品質を維持する1ビットHy3量子化のユーザーテスト https://korshunov.ai/ar/article/12223-khtbrt-lmstkhdmyn-lkmy-hy3-bbt-whd-thfz-l-ljwd-nd-89-jyjbyt/korshunov.aiar2026-07-16T01:04:00.101002+00:00اختبارات المستخدمين لكمية Hy3 ببت واحد تحافظ على الجودة عند 89 جيجابايت https://korshunov.ai/id/article/12223-pengujian-pengguna-kuantisasi-hy3-1-bit-mempertahankan-kualitas-pada-89-gb/korshunov.aiid2026-07-16T01:04:00.101002+00:00Pengujian pengguna kuantisasi Hy3 1-bit mempertahankan kualitas pada 89 GB https://korshunov.ai/fr/article/12223-tests-utilisateurs-de-la-quantisation-hy3-1-bit-conservant-la-qualite-a-89-go/korshunov.aifr2026-07-16T01:04:00.101002+00:00Tests utilisateurs de la quantisation Hy3 1-bit conservant la qualité à 89 Go https://korshunov.ai/en/article/12223-user-tests-1-bit-hy3-quantization-retaining-quality-at-89-gb/korshunov.aien2026-07-16T01:04:00.101002+00:00User tests 1-bit Hy3 quantization retaining quality at 89 GB https://korshunov.ai/ko/article/12223-89-gbeseo-pumjileul-yujihaneun-1biteu-hy3-yangjahwa-sayongja-teseuteu/korshunov.aiko2026-07-16T01:04:00.101002+00:0089 GB에서 품질을 유지하는 1비트 Hy3 양자화 사용자 테스트 https://korshunov.ai/ru/article/12223-testy-polzovatelei-1-bitnoe-kvantovanie-hy3-sokhraniaet-kachestvo-pri-razmere-89/korshunov.airu2026-07-16T01:04:00.101002+00:00Тесты пользователей: 1-битное квантование Hy3 сохраняет качество при размере 89 ГБ https://korshunov.ai/ar/article/12224-kymy-k3-yulmwh-lsdrh-zhr-blf-l-fy-lsh-bsm-kwd-kivine/korshunov.aiar2026-07-16T00:57:53.265603+00:00كيمي k3 يُلمّح لإصداره، ظهر بالفعل في الساحة باسم كود kivine https://korshunov.ai/id/article/12224-kimi-k3-diisyaratkan-akan-dirilis-sudah-ada-di-arena-sebagai-kivine/korshunov.aiid2026-07-16T00:57:53.265603+00:00Kimi k3 diisyaratkan akan dirilis, sudah ada di arena sebagai kivine https://korshunov.ai/fr/article/12224-kimi-k3-annonce-pour-sa-sortie-deja-present-sur-l-arene-sous-le-nom-de-code/korshunov.aifr2026-07-16T00:57:53.265603+00:00Kimi k3 annoncé pour sa sortie, déjà présent sur l'arène sous le nom de code kivine https://korshunov.ai/ko/article/12224-kimi-k3-culsi-yego-imi-kivinero-arenae-deungjang/korshunov.aiko2026-07-16T00:57:53.265603+00:00Kimi k3 출시 예고, 이미 kivine로 아레나에 등장 https://korshunov.ai/ja/article/12224-kimi-k3noriri-suyu-gao-sutenikivinetoshiteari-nanideng-chang/korshunov.aija2026-07-16T00:57:53.265603+00:00Kimi k3のリリース予告、すでにkivineとしてアリーナに登場 https://korshunov.ai/hi/article/12224-kimi-k3-ke-riliij-kaa-snket-kivine-ke-ruup-men-phle-se-hii-arenaa-pr/korshunov.aihi2026-07-16T00:57:53.265603+00:00Kimi k3 के रिलीज का संकेत, kivine के रूप में पहले से ही अरेना पर https://korshunov.ai/pt-br/article/12224-kimi-k3-anunciado-para-lancamento-ja-na-arena-como-kivine/korshunov.aipt2026-07-16T00:57:53.265603+00:00Kimi k3 anunciado para lançamento, já na arena como kivine https://korshunov.ai/zh-hans/article/12224-kimi-k3-yu-gao-fa-bu-yi-yi-kivine-dai-hao-chu-xian-zai-jing-ji-chang/korshunov.aizh-cn2026-07-16T00:57:53.265603+00:00Kimi k3 预告发布，已以 kivine 代号出现在竞技场 https://korshunov.ai/es/article/12224-kimi-k3-se-anuncia-para-su-lanzamiento-ya-en-la-arena-como-kivine/korshunov.aies2026-07-16T00:57:53.265603+00:00Kimi k3 se anuncia para su lanzamiento, ya en la arena como kivine https://korshunov.ai/ru/article/12224-kimi-k3-anonsirovan-k-vypusku-uzhe-dostupen-na-arene-pod-psevdonimom-kivine/korshunov.airu2026-07-16T00:57:53.265603+00:00Kimi k3 анонсирован к выпуску, уже доступен на арене под псевдонимом kivine https://korshunov.ai/en/article/12224-kimi-k3-teased-for-release-already-on-arena-as-kivine/korshunov.aien2026-07-16T00:57:53.265603+00:00Kimi k3 teased for release, already on arena as kivine https://korshunov.ai/ko/article/12225-xai-peuraibaeglash-hu-grok-build-kodeubeiseu-opeunsoseuhwa/korshunov.aiko2026-07-16T00:48:14.114767+00:00xAI, 프라이백lash 후 Grok Build 코드베이스 오픈소스화 https://korshunov.ai/ru/article/12225-xai-otkryl-iskhodnyi-kod-grok-build-na-fone-skandala-s-konfidentsialnostiu/korshunov.airu2026-07-16T00:48:14.114767+00:00xAI открыл исходный код Grok Build на фоне скандала с конфиденциальностью https://korshunov.ai/en/article/12225-xai-open-sources-grok-build-codebase-following-privacy-backlash/korshunov.aien2026-07-16T00:48:14.114767+00:00xAI open sources Grok Build codebase following privacy backlash https://korshunov.ai/pt-br/article/12225-xai-abre-o-codigo-fonte-do-grok-build-apos-controversia-de-privacidade/korshunov.aipt2026-07-16T00:48:14.114767+00:00xAI abre o código-fonte do Grok Build após controvérsia de privacidade https://korshunov.ai/ja/article/12225-xai-huraihashi-wen-ti-woshou-ketegrok-buildnoko-tohe-suwoo-hunso-suhua/korshunov.aija2026-07-16T00:48:14.114767+00:00xAI、プライバシー問題を受けてGrok Buildのコードベースをオープンソース化 https://korshunov.ai/fr/article/12225-xai-open-source-le-codebase-de-grok-build-suite-a-une-controverse-sur-la/korshunov.aifr2026-07-16T00:48:14.114767+00:00xAI open-source le codebase de Grok Build suite à une controverse sur la confidentialité https://korshunov.ai/hi/article/12225-gopniiytaa-vivaad-ke-baad-xai-ne-grok-build-koddbes-opn-sors-kiyaa/korshunov.aihi2026-07-16T00:48:14.114767+00:00गोपनीयता विवाद के बाद xAI ने Grok Build कोडबेस ओपन सोर्स किया https://korshunov.ai/zh-hans/article/12225-xaizai-yin-si-zheng-yi-hou-kai-yuan-grok-builddai-ma-ku/korshunov.aizh-cn2026-07-16T00:48:14.114767+00:00xAI在隐私争议后开源Grok Build代码库 https://korshunov.ai/es/article/12225-xai-abre-el-codigo-fuente-de-grok-build-tras-la-controversia-por-privacidad/korshunov.aies2026-07-16T00:48:14.114767+00:00xAI abre el código fuente de Grok Build tras la controversia por privacidad https://korshunov.ai/ar/article/12225-xai-tfth-msdr-kwd-grok-build-b-d-rd-f-l-nyf-bshn-lkhswsy/korshunov.aiar2026-07-16T00:48:14.114767+00:00xAI تفتح مصدر كود Grok Build بعد رد فعل عنيف بشأن الخصوصية https://korshunov.ai/id/article/12225-xai-membuka-sumber-basis-kode-grok-build-setelah-kecaman-privasi/korshunov.aiid2026-07-16T00:48:14.114767+00:00xAI membuka sumber basis kode Grok Build setelah kecaman privasi https://korshunov.ai/id/article/12220-claude-code-v2-1-211-memperbaiki-pratinjau-izin-hang-sesi-dan-penagihan-caching/korshunov.aiid2026-07-16T00:40:25.557142+00:00Claude Code v2.1.211 memperbaiki pratinjau izin, hang sesi, dan penagihan caching prompt https://korshunov.ai/fr/article/12220-claude-code-v2-1-211-corrige-les-apercus-de-permissions-les-blocages-de-session/korshunov.aifr2026-07-16T00:40:25.557142+00:00Claude Code v2.1.211 corrige les aperçus de permissions, les blocages de session et la facturation du cache de prompts https://korshunov.ai/ko/article/12220-claude-code-v2-1-211i-gweonhan-miribogi-sesyeon-jungdan-mic-peurompeuteu-kaesing/korshunov.aiko2026-07-16T00:40:25.557142+00:00Claude Code v2.1.211이 권한 미리보기, 세션 중단 및 프롬프트 캐싱 과금을 수정함 https://korshunov.ai/ar/article/12220-yslh-claude-code-v2-1-211-m-ynt-ldhwnt-t-lyq-ljlst-wfwtr-ltkhzyn-lmwqt-lltlbt/korshunov.aiar2026-07-16T00:40:25.557142+00:00يصلح Claude Code v2.1.211 معاينات الأذونات، تعليق الجلسات، وفوترة التخزين المؤقت للطلبات https://korshunov.ai/pt-br/article/12220-claude-code-v2-1-211-corrige-visualizacoes-de-permissao-travamentos-de-sessao-e/korshunov.aipt2026-07-16T00:40:25.557142+00:00Claude Code v2.1.211 corrige visualizações de permissão, travamentos de sessão e cobrança de cache de prompt https://korshunov.ai/zh-hans/article/12220-claude-code-v2-1-211-xiu-fu-quan-xian-yu-lan-hui-hua-gua-qi-he-ti-shi-huan-cun/korshunov.aizh-cn2026-07-16T00:40:25.557142+00:00Claude Code v2.1.211 修复权限预览、会话挂起和提示缓存计费问题 https://korshunov.ai/en/article/12220-claude-code-v2-1-211-fixes-permission-previews-session-hangs-and-prompt-caching/korshunov.aien2026-07-16T00:40:25.557142+00:00Claude Code v2.1.211 fixes permission previews, session hangs, and prompt-caching billing https://korshunov.ai/ru/article/12220-claude-code-v2-1-211-ispravliaet-predvaritelnyi-prosmotr-razreshenii-zavisaniia/korshunov.airu2026-07-16T00:40:25.557142+00:00Claude Code v2.1.211 исправляет предварительный просмотр разрешений, зависания сессий и биллинг кэширования промптов https://korshunov.ai/es/article/12220-claude-code-v2-1-211-corrige-vistas-previas-de-permisos-bloqueos-de-sesion-y-de/korshunov.aies2026-07-16T00:40:25.557142+00:00Claude Code v2.1.211 corrige vistas previas de permisos, bloqueos de sesión y facturación de caché de prompts https://korshunov.ai/hi/article/12220-claude-code-v2-1-211-anumti-puurvaavlokn-str-hng-aur-pronmptt-kaishing-biling-ko/korshunov.aihi2026-07-16T00:40:25.557142+00:00Claude Code v2.1.211 अनुमति पूर्वावलोकन, सत्र हंग और प्रॉम्प्ट-कैशिंग बिलिंग को ठीक करता है https://korshunov.ai/ja/article/12220-claude-code-v2-1-211-kaquan-xian-hurehiyu-setsushiyonnohanku-jin-woxiu-zheng/korshunov.aija2026-07-16T00:40:25.557142+00:00Claude Code v2.1.211 が権限プレビュー、セッションのハング、プロンプトキャッシングの課金を修正 https://korshunov.ai/ko/article/12226-spacexai-grok-build-koding-eijeonteu-mic-tui-opeunsoseuhwa/korshunov.aiko2026-07-16T00:33:01.232347+00:00SpaceXAI, Grok Build 코딩 에이전트 및 TUI 오픈소스화 https://korshunov.ai/ja/article/12226-spacexaikagrok-buildnoko-teinkue-shientototuiwoo-hunso-suhua/korshunov.aija2026-07-16T00:33:01.232347+00:00SpaceXAIがGrok BuildのコーディングエージェントとTUIをオープンソース化 https://korshunov.ai/id/article/12226-spacexai-membuka-sumber-agen-pemrograman-grok-build-dan-tui/korshunov.aiid2026-07-16T00:33:01.232347+00:00SpaceXAI membuka sumber agen pemrograman Grok Build dan TUI https://korshunov.ai/ar/article/12226-spacexai-tfth-msdr-wkyl-lbrmj-grok-build-wwjh-tui/korshunov.aiar2026-07-16T00:33:01.232347+00:00SpaceXAI تفتح مصدر وكيل البرمجة Grok Build وواجهة TUI https://korshunov.ai/pt-br/article/12226-spacexai-abre-o-codigo-fonte-do-grok-build-agente-de-codificacao-e-tui/korshunov.aipt2026-07-16T00:33:01.232347+00:00SpaceXAI abre o código-fonte do Grok Build, agente de codificação e TUI https://korshunov.ai/hi/article/12226-spacexai-ne-grok-build-kodding-ejentt-aur-tui-kaa-opn-sors-kiyaa/korshunov.aihi2026-07-16T00:33:01.232347+00:00SpaceXAI ने Grok Build कोडिंग एजेंट और TUI का ओपन-सोर्स किया https://korshunov.ai/en/article/12226-spacexai-open-sources-grok-build-coding-agent-and-tui/korshunov.aien2026-07-16T00:33:01.232347+00:00SpaceXAI open-sources Grok Build coding agent and TUI https://korshunov.ai/ru/article/12226-spacexai-otkryla-iskhodnyi-kod-grok-build-agenta-dlia-programmirovaniia-i-tui/korshunov.airu2026-07-16T00:33:01.232347+00:00SpaceXAI открыла исходный код Grok Build — агента для программирования и TUI https://korshunov.ai/fr/article/12226-spacexai-open-source-l-agent-de-codage-grok-build-et-l-interface-tui/korshunov.aifr2026-07-16T00:33:01.232347+00:00SpaceXAI open-source l'agent de codage Grok Build et l'interface TUI https://korshunov.ai/es/article/12226-spacexai-abre-el-codigo-fuente-de-grok-build-agente-de-codificacion-y-tui/korshunov.aies2026-07-16T00:33:01.232347+00:00SpaceXAI abre el código fuente de Grok Build, agente de codificación y TUI https://korshunov.ai/zh-hans/article/12226-spacexai-kai-yuan-grok-build-bian-ma-dai-li-he-tui/korshunov.aizh-cn2026-07-16T00:33:01.232347+00:00SpaceXAI 开源 Grok Build 编码代理和 TUI https://korshunov.ai/ko/article/12140-pope-pyeongganeun-donggyeoldoen-sohyeong-kodeu-modeleseo-oryu-naeyongi-jace/korshunov.aiko2026-07-16T00:24:21.983388+00:00PoPE 평가는 동결된 소형 코드 모델에서 오류 내용이 자체 수리에 도움이 된다는 증거가 없음을 발견 https://korshunov.ai/fr/article/12140-l-evaluation-pope-ne-trouve-aucune-preuve-que-le-contenu-de-l-erreur-aide-a-l-de/korshunov.aifr2026-07-16T00:24:21.983388+00:00L'évaluation PoPE ne trouve aucune preuve que le contenu de l'erreur aide à l'auto-réparation dans les petits modèles de code gelés https://korshunov.ai/ja/article/12140-popeping-jia-ha-dong-jie-saretaxiao-gui-mo-ko-tomoteruniokeruera-nei-rong-kazi/korshunov.aija2026-07-16T00:24:21.983388+00:00PoPE評価は、凍結された小規模コードモデルにおけるエラー内容が自己修復に寄与する証拠がないことを発見 https://korshunov.ai/hi/article/12140-pope-muulyaankn-men-koii-sbuut-nhiin-miltaa-ki-trutti-saamgrii-jmii-huii-chotte/korshunov.aihi2026-07-16T00:24:21.983388+00:00PoPE मूल्यांकन में कोई सबूत नहीं मिलता कि त्रुटि सामग्री जमी हुई छोटे कोड मॉडलों में स्व-मरम्मत में सहायक है https://korshunov.ai/zh-hans/article/12140-popeping-gu-wei-fa-xian-cuo-wu-nei-rong-you-zhu-yu-dong-jie-xiao-xing-dai-ma-mo/korshunov.aizh-cn2026-07-16T00:24:21.983388+00:00PoPE评估未发现错误内容有助于冻结小型代码模型自我修复的证据 https://korshunov.ai/en/article/12140-pope-evaluation-finds-no-evidence-that-error-content-aids-self-repair-in-frozen/korshunov.aien2026-07-16T00:24:21.983388+00:00PoPE evaluation finds no evidence that error content aids self-repair in frozen small code models https://korshunov.ai/id/article/12140-evaluasi-pope-tidak-menemukan-bukti-bahwa-konten-kesalahan-membantu-perbaikan/korshunov.aiid2026-07-16T00:24:21.983388+00:00Evaluasi PoPE tidak menemukan bukti bahwa konten kesalahan membantu perbaikan mandiri pada model kode kecil yang dibekukan https://korshunov.ai/ar/article/12140-tqyym-pope-l-yjd-y-dlyl-l-n-mhtw-lkht-ys-d-fy-lslh-ldhty-lnmdhj-lshyfr-lsgyr/korshunov.aiar2026-07-16T00:24:21.983388+00:00تقييم PoPE لا يجد أي دليل على أن محتوى الخطأ يساعد في الإصلاح الذاتي لنماذج الشيفرة الصغيرة المتجمدة https://korshunov.ai/ru/article/12140-otsenka-pope-ne-nakhodit-dokazatelstv-togo-chto-soderzhanie-oshibok-pomogaet/korshunov.airu2026-07-16T00:24:21.983388+00:00Оценка PoPE не находит доказательств того, что содержание ошибок помогает самовосстановлению замороженных малых моделей кода https://korshunov.ai/es/article/12140-la-evaluacion-pope-no-encuentra-evidencia-de-que-el-contenido-del-error-ayude-a/korshunov.aies2026-07-16T00:24:21.983388+00:00La evaluación PoPE no encuentra evidencia de que el contenido del error ayude a la autocorrección en modelos pequeños de código congelados https://korshunov.ai/pt-br/article/12140-avaliacao-pope-nao-encontra-evidencias-de-que-o-conteudo-do-erro-auxilie-na-de/korshunov.aipt2026-07-16T00:24:21.983388+00:00Avaliação PoPE não encontra evidências de que o conteúdo do erro auxilie na autorreparação de pequenos modelos de código congelados https://korshunov.ai/pt-br/article/12139-filtro-de-fluxo-controlado-por-conjunto-encf-para-assimilacao-implicita-de-dados/korshunov.aipt2026-07-16T00:14:43.447222+00:00Filtro de Fluxo Controlado por Conjunto (EnCF) para assimilação implícita de dados https://korshunov.ai/hi/article/12139-aspsstt-ddettaa-esiimileshn-ke-lie-ensenbl-knttroldd-phlo-philttr-encf/korshunov.aihi2026-07-16T00:14:43.447222+00:00अस्पष्ट डेटा एसीमिलेशन के लिए एनसेंबल कंट्रोल्ड-फ्लो फ़िल्टर (EnCF) https://korshunov.ai/es/article/12139-filtro-de-flujo-controlado-por-conjunto-encf-para-asimilacion-implicita-de-datos/korshunov.aies2026-07-16T00:14:43.447222+00:00Filtro de flujo controlado por conjunto (EnCF) para asimilación implícita de datos https://korshunov.ai/ja/article/12139-an-mo-de-te-tatong-hua-notamenoansanhuruzhi-yu-huro-huiruta-encf/korshunov.aija2026-07-16T00:14:43.447222+00:00暗黙的データ同化のためのアンサンブル制御フローフィルタ (EnCF) https://korshunov.ai/ko/article/12139-amsijeog-deiteo-donghwareul-wihan-angsangbeul-jeeo-heureum-pilteo-encf/korshunov.aiko2026-07-16T00:14:43.447222+00:00암시적 데이터 동화를 위한 앙상블 제어 흐름 필터 (EnCF) https://korshunov.ai/zh-hans/article/12139-yong-yu-yin-shi-shu-ju-tong-hua-de-ji-cheng-kong-zhi-liu-lu-bo-qi-encf/korshunov.aizh-cn2026-07-16T00:14:43.447222+00:00用于隐式数据同化的集成控制流滤波器 (EnCF) https://korshunov.ai/ru/article/12139-filtr-potoka-s-upravleniem-ansamblem-encf-dlia-neiavnogo-assimilirovaniia/korshunov.airu2026-07-16T00:14:43.447222+00:00Фильтр потока с управлением ансамблем (EnCF) для неявного ассимилирования данных https://korshunov.ai/fr/article/12139-filtre-a-flot-controle-par-ensemble-encf-pour-l-assimilation-de-donnees/korshunov.aifr2026-07-16T00:14:43.447222+00:00Filtre à flot contrôlé par ensemble (EnCF) pour l'assimilation de données implicite https://korshunov.ai/id/article/12139-filter-aliran-terkendali-ensemble-encf-untuk-asimilasi-data-implisit/korshunov.aiid2026-07-16T00:14:43.447222+00:00Filter Aliran Terkendali Ensemble (EnCF) untuk asimilasi data implisit https://korshunov.ai/ar/article/12139-mrshh-ltdfq-lmthkm-bh-bltjm-encf-llmhdh-ldmny-llbynt/korshunov.aiar2026-07-16T00:14:43.447222+00:00مرشح التدفق المتحكم به بالتجمع (EnCF) للمحاذاة الضمنية للبيانات https://korshunov.ai/en/article/12139-ensemble-controlled-flow-filter-encf-for-implicit-data-assimilation/korshunov.aien2026-07-16T00:14:43.447222+00:00Ensemble Controlled-flow Filter (EnCF) for implicit data assimilation https://korshunov.ai/pt-br/article/12091-formalanalyticgeo-gera-problemas-de-geometria-analitica-multimodal-via-pipeline/korshunov.aipt2026-07-16T00:07:11.429674+00:00FormalAnalyticGeo gera problemas de geometria analítica multimodal via pipeline neural-simbólico https://korshunov.ai/zh-hans/article/12091-formalanalyticgeotong-guo-shen-jing-fu-hao-liu-shui-xian-sheng-cheng-duo-mo-tai/korshunov.aizh-cn2026-07-16T00:07:11.429674+00:00FormalAnalyticGeo通过神经符号流水线生成多模态解析几何问题 https://korshunov.ai/hi/article/12091-formalanalyticgeo-nyuuro-sinbolik-paaiplaain-ke-maadhym-se-bhuaayaamii-jyaamiti/korshunov.aihi2026-07-16T00:07:11.429674+00:00FormalAnalyticGeo न्यूरो-सिंबोलिक पाइपलाइन के माध्यम से बहुआयामी विश्लेषणात्मक ज्यामिति की समस्याएं उत्पन्न करता है https://korshunov.ai/es/article/12091-formalanalyticgeo-genera-problemas-de-geometria-analitica-multimodal-mediante-un/korshunov.aies2026-07-16T00:07:11.429674+00:00FormalAnalyticGeo genera problemas de geometría analítica multimodal mediante un pipeline neural-simbólico https://korshunov.ai/ru/article/12091-formalanalyticgeo-generiruet-multimodalnye-zadachi-analiticheskoi-geometrii-s/korshunov.airu2026-07-16T00:07:11.429674+00:00FormalAnalyticGeo генерирует мультимодальные задачи аналитической геометрии с помощью нейросимволического конвейера https://korshunov.ai/ja/article/12091-formalanalyticgeohaniyu-raru-shinhoritsukuhaihurainwotong-shiteduo-mo-tarujie-xi/korshunov.aija2026-07-16T00:07:11.429674+00:00FormalAnalyticGeoはニューラル・シンボリックパイプラインを通じて多モーダル解析幾何学問題を生成する https://korshunov.ai/en/article/12091-formalanalyticgeo-generates-multimodal-analytic-geometry-problems-via-neural/korshunov.aien2026-07-16T00:07:11.429674+00:00FormalAnalyticGeo generates multimodal analytic geometry problems via neural-symbolic pipeline https://korshunov.ai/id/article/12091-formalanalyticgeo-menghasilkan-masalah-geometri-analitik-multimoda-melalui/korshunov.aiid2026-07-16T00:07:11.429674+00:00FormalAnalyticGeo menghasilkan masalah geometri analitik multimoda melalui pipeline neural-symbolic https://korshunov.ai/ko/article/12091-formalanalyticgeoneun-singyeong-giho-paipeuraineul-tonghae-dajung-modal-haeseog/korshunov.aiko2026-07-16T00:07:11.429674+00:00FormalAnalyticGeo는 신경-기호 파이프라인을 통해 다중 모달 해석 기하학 문제를 생성합니다 https://korshunov.ai/fr/article/12091-formalanalyticgeo-genere-des-problemes-de-geometrie-analytique-multimodale-via/korshunov.aifr2026-07-16T00:07:11.429674+00:00FormalAnalyticGeo génère des problèmes de géométrie analytique multimodale via un pipeline neuro-symbolique https://korshunov.ai/ar/article/12091-formalanalyticgeo-ywld-msyl-lhnds-lthlyly-mt-dd-lwsyt-br-kht-nbyb-rmzy-sby/korshunov.aiar2026-07-16T00:07:11.429674+00:00FormalAnalyticGeo يولد مسائل الهندسة التحليلية متعددة الوسائط عبر خط أنابيب رمزي-عصبي https://korshunov.ai/ja/article/12090-di-kang-togeng-xin-insenteihushi-he-llmnotamenofan-shi-shi-de-reho-tozuo-biao/korshunov.aija2026-07-15T23:58:11.507569+00:00抵抗と更新：インセンティブ適合LLMのための反事実的レポート座標 https://korshunov.ai/hi/article/12090-prtirodh-kren-aur-adytn-kren-prernnaa-sngt-llms-ke-lie-prtigaamii-riportt/korshunov.aihi2026-07-15T23:58:11.507569+00:00प्रतिरोध करें और अद्यतन करें: प्रेरणा-संगत LLMs के लिए प्रतिगामी रिपोर्ट निर्देशांक https://korshunov.ai/pt-br/article/12090-resistir-e-atualizar-coordenadas-de-relatos-contrafatuais-para-llms-compativeis/korshunov.aipt2026-07-15T23:58:11.507569+00:00Resistir e Atualizar: Coordenadas de Relatos Contrafatuais para LLMs Compatíveis com Incentivos https://korshunov.ai/zh-hans/article/12090-di-kang-yu-geng-xin-yong-yu-ji-li-jian-rong-llm-de-fan-shi-shi-bao-gao-zuo-biao/korshunov.aizh-cn2026-07-15T23:58:11.507569+00:00抵抗与更新：用于激励兼容 LLM 的反事实报告坐标 https://korshunov.ai/en/article/12090-resist-and-update-counterfactual-report-coordinates-for-incentive-compatible/korshunov.aien2026-07-15T23:58:11.507569+00:00Resist and Update: Counterfactual Report Coordinates for Incentive-Compatible LLMs https://korshunov.ai/ru/article/12090-soprotivliaites-i-obnovliaites-koordinaty-kontrfakticheskikh-otchetov-dlia-llm/korshunov.airu2026-07-15T23:58:11.507569+00:00Сопротивляйтесь и обновляйтесь: координаты контрфактических отчетов для мотивационно-совместимых LLM https://korshunov.ai/es/article/12090-resistir-y-actualizar-coordenadas-de-informes-contrafacticos-para-llms-con/korshunov.aies2026-07-15T23:58:11.507569+00:00Resistir y actualizar: coordenadas de informes contrafácticos para LLMs compatibles con incentivos https://korshunov.ai/ko/article/12090-jeohang-mic-eobdeiteu-insentibeu-hohwan-llmeul-wihan-bansasiljeog-bogo-jwapyo/korshunov.aiko2026-07-15T23:58:11.507569+00:00저항 및 업데이트: 인센티브 호환 LLM을 위한 반사실적 보고 좌표 https://korshunov.ai/ar/article/12090-lmqwm-wlthdyth-hdthyt-ltqryr-lmdd-llwq-lnmdhj-llgt-lkbyr-lmtwfq-m-lhwfz/korshunov.aiar2026-07-15T23:58:11.507569+00:00المقاومة والتحديث: إحداثيات التقارير المضادة للواقع لنماذج اللغات الكبيرة المتوافقة مع الحوافز https://korshunov.ai/fr/article/12090-resister-et-mettre-a-jour-coordonnees-de-rapport-contrefactuel-pour-des-llm-avec/korshunov.aifr2026-07-15T23:58:11.507569+00:00Résister et Mettre à Jour : Coordonnées de rapport contrefactuel pour des LLM compatibles avec les incitations https://korshunov.ai/id/article/12090-menolak-dan-memperbarui-koordinat-laporan-kontrfaktual-untuk-llm-yang-kompatibel/korshunov.aiid2026-07-15T23:58:11.507569+00:00Menolak dan Memperbarui: Koordinat Laporan Kontrfaktual untuk LLM yang Kompatibel dengan Insentif https://korshunov.ai/ru/article/12194-nvidia-cuda-13-3-dobavliaet-nativnuiu-podderzhku-umnozheniia-bez-perenosa/korshunov.airu2026-07-15T23:48:11.408393+00:00NVIDIA CUDA 13.3 добавляет нативную поддержку умножения без переноса https://korshunov.ai/pt-br/article/12194-nvidia-cuda-13-3-adiciona-suporte-nativo-para-multiplicacao-sem-carry/korshunov.aipt2026-07-15T23:48:11.408393+00:00NVIDIA CUDA 13.3 adiciona suporte nativo para multiplicação sem carry https://korshunov.ai/hi/article/12194-nvidia-cuda-13-3-men-nettiv-kairiiles-mlttiiplikeshn-smrthn-joddaa-gyaa/korshunov.aihi2026-07-15T23:48:11.408393+00:00NVIDIA CUDA 13.3 में नेटिव कैरीलेस मल्टीप्लिकेशन समर्थन जोड़ा गया https://korshunov.ai/ja/article/12194-nvidia-cuda-13-3-kaneiteihuno-carryless-multiplication-saho-towozhui-jia/korshunov.aija2026-07-15T23:48:11.408393+00:00NVIDIA CUDA 13.3 がネイティブの carryless multiplication サポートを追加 https://korshunov.ai/ko/article/12194-nvidia-cuda-13-3-neitibeu-carryless-multiplication-jiweon-cuga/korshunov.aiko2026-07-15T23:48:11.408393+00:00NVIDIA CUDA 13.3, 네이티브 carryless multiplication 지원 추가 https://korshunov.ai/es/article/12194-nvidia-cuda-13-3-anade-soporte-nativo-para-multiplicacion-sin-acarreo/korshunov.aies2026-07-15T23:48:11.408393+00:00NVIDIA CUDA 13.3 añade soporte nativo para multiplicación sin acarreo https://korshunov.ai/fr/article/12194-nvidia-cuda-13-3-ajoute-le-support-natif-de-la-multiplication-carryless/korshunov.aifr2026-07-15T23:48:11.408393+00:00NVIDIA CUDA 13.3 ajoute le support natif de la multiplication carryless https://korshunov.ai/zh-hans/article/12194-nvidia-cuda-13-3-tian-jia-yuan-sheng-wu-jin-wei-cheng-fa-zhi-chi/korshunov.aizh-cn2026-07-15T23:48:11.408393+00:00NVIDIA CUDA 13.3 添加原生无进位乘法支持 https://korshunov.ai/en/article/12194-nvidia-cuda-13-3-adds-native-carryless-multiplication-support/korshunov.aien2026-07-15T23:48:11.408393+00:00NVIDIA CUDA 13.3 adds native carryless multiplication support https://korshunov.ai/id/article/12194-nvidia-cuda-13-3-menambahkan-dukungan-perkalian-carryless-native/korshunov.aiid2026-07-15T23:48:11.408393+00:00NVIDIA CUDA 13.3 menambahkan dukungan perkalian carryless native https://korshunov.ai/ar/article/12194-nvidia-cuda-13-3-tdyf-d-m-ldrb-lkhly-mn-lhml-carryless-multiplication-bshkl-sly/korshunov.aiar2026-07-15T23:48:11.408393+00:00NVIDIA CUDA 13.3 تضيف دعم الضرب الخالي من الحمل (carryless multiplication) بشكل أصلي https://korshunov.ai/zh-hans/article/12183-yong-hu-zai-da-zai-graphene-os-de-android-shang-yun-xing-hermes-hou-duan-wei-cpp/korshunov.aizh-cn2026-07-15T23:43:40.584115+00:00用户在搭载 Graphene OS 的 Android 上运行 Hermes，后端为远程 Llama.cpp https://korshunov.ai/es/article/12183-un-usuario-ejecuta-hermes-en-android-con-graphene-os-y-backend-remoto-de-llama/korshunov.aies2026-07-15T23:43:40.584115+00:00Un usuario ejecuta Hermes en Android con Graphene OS y backend remoto de Llama.cpp https://korshunov.ai/en/article/12183-user-runs-hermes-on-android-with-graphene-os-and-remote-llama-cpp-backend/korshunov.aien2026-07-15T23:43:40.584115+00:00User runs Hermes on Android with Graphene OS and remote Llama.cpp backend https://korshunov.ai/ru/article/12183-polzovatel-zapuskaet-hermes-na-android-s-graphene-os-i-udalennym-bekendom-llama/korshunov.airu2026-07-15T23:43:40.584115+00:00Пользователь запускает Hermes на Android с Graphene OS и удалённым бэкендом Llama.cpp https://korshunov.ai/hi/article/12183-upyogkrtaa-graphene-os-aur-rimott-llama-cpp-baikendd-ke-saath-android-pr-hermes/korshunov.aihi2026-07-15T23:43:40.584115+00:00उपयोगकर्ता Graphene OS और रिमोट Llama.cpp बैकएंड के साथ Android पर Hermes चला रहा है https://korshunov.ai/ja/article/12183-graphene-osshang-noandroidtehermeswoshi-xing-shi-rimo-tollama-yong/korshunov.aija2026-07-15T23:43:40.584115+00:00Graphene OS上のAndroidでHermesを実行し、リモートLlama.cppバックエンドを使用 https://korshunov.ai/ko/article/12183-graphene-osga-seolcidoen-androideseo-hermes-silhaeng-mic-weongyeog-llama-cpp/korshunov.aiko2026-07-15T23:43:40.584115+00:00Graphene OS가 설치된 Android에서 Hermes 실행 및 원격 Llama.cpp 백엔드 사용 https://korshunov.ai/pt-br/article/12183-usuario-executa-hermes-no-android-com-graphene-os-e-backend-remoto-do-llama-cpp/korshunov.aipt2026-07-15T23:43:40.584115+00:00Usuário executa Hermes no Android com Graphene OS e backend remoto do Llama.cpp https://korshunov.ai/fr/article/12183-un-utilisateur-execute-hermes-sur-android-avec-graphene-os-et-un-backend-llama/korshunov.aifr2026-07-15T23:43:40.584115+00:00Un utilisateur exécute Hermes sur Android avec Graphene OS et un backend Llama.cpp distant https://korshunov.ai/ar/article/12183-mstkhdm-yshgl-hermes-l-ndrwyd-m-graphene-os-wkhdm-llama-cpp-n-bu-d/korshunov.aiar2026-07-15T23:43:40.584115+00:00مستخدم يشغل Hermes على أندرويد مع Graphene OS وخادم Llama.cpp عن بُعد https://korshunov.ai/id/article/12183-pengguna-menjalankan-hermes-di-android-dengan-graphene-os-dan-backend-llama-cpp/korshunov.aiid2026-07-15T23:43:40.584115+00:00Pengguna menjalankan Hermes di Android dengan Graphene OS dan backend Llama.cpp jarak jauh https://korshunov.ai/ko/article/12181-dogil-ai-keonsosieomi-opeun-30b-soofi-s-model-culsi/korshunov.aiko2026-07-15T23:36:09.420028+00:00독일 AI 컨소시엄이 오픈 30B Soofi S 모델 출시 https://korshunov.ai/ja/article/12181-toitsunoaikonso-shiamukao-hunna30b-soofi-smoteruworiri-su/korshunov.aija2026-07-15T23:36:09.420028+00:00ドイツのAIコンソーシアムがオープンな30B Soofi Sモデルをリリース https://korshunov.ai/en/article/12181-german-ai-consortium-releases-open-30b-soofi-s-model/korshunov.aien2026-07-15T23:36:09.420028+00:00German AI consortium releases open 30B Soofi S model https://korshunov.ai/zh-hans/article/12181-de-guo-ailian-meng-fa-bu-kai-yuan-30b-soofi-smo-xing/korshunov.aizh-cn2026-07-15T23:36:09.420028+00:00德国AI联盟发布开源30B Soofi S模型 https://korshunov.ai/fr/article/12181-un-consortium-d-ia-allemand-publie-le-modele-soofi-s-ouvert-de-30-milliards-de/korshunov.aifr2026-07-15T23:36:09.420028+00:00Un consortium d'IA allemand publie le modèle Soofi S ouvert de 30 milliards de paramètres https://korshunov.ai/ru/article/12181-nemetskii-konsortsium-po-ii-vypustil-otkrytuiu-model-soofi-s-na-30b/korshunov.airu2026-07-15T23:36:09.420028+00:00Немецкий консорциум по ИИ выпустил открытую модель Soofi S на 30B https://korshunov.ai/es/article/12181-consorcio-aleman-de-ia-lanza-el-modelo-soofi-s-abierto-de-30b/korshunov.aies2026-07-15T23:36:09.420028+00:00Consorcio alemán de IA lanza el modelo Soofi S abierto de 30B https://korshunov.ai/pt-br/article/12181-consorcio-de-ia-alemao-lanca-modelo-soofi-s-aberto-de-30b/korshunov.aipt2026-07-15T23:36:09.420028+00:00Consórcio de IA alemão lança modelo Soofi S aberto de 30B https://korshunov.ai/ar/article/12181-thd-ldhk-lstn-y-llmny-ytlq-nmwdhj-soofi-s-lmftwh-b-30-mlyr-m-lm/korshunov.aiar2026-07-15T23:36:09.420028+00:00اتحاد الذكاء الاصطناعي الألماني يطلق نموذج Soofi S المفتوح بـ 30 مليار معلمة https://korshunov.ai/hi/article/12181-jrmn-ai-knsorttiym-ne-opn-30b-soofi-s-monddl-jaarii-kiyaa/korshunov.aihi2026-07-15T23:36:09.420028+00:00जर्मन AI कंसोर्टियम ने ओपन 30B Soofi S मॉडल जारी किया https://korshunov.ai/id/article/12181-konsorsium-ai-jerman-merilis-model-soofi-s-terbuka-30b/korshunov.aiid2026-07-15T23:36:09.420028+00:00Konsorsium AI Jerman merilis model Soofi S terbuka 30B https://korshunov.ai/fr/article/12217-un-utilisateur-demande-une-comparaison-entre-antirez-ds4-et-unsloth-gguf-en-2/korshunov.aifr2026-07-15T23:30:55.274931+00:00Un utilisateur demande une comparaison entre Antirez DS4 et Unsloth GGUF en quantification 2-bit https://korshunov.ai/ko/article/12217-sayongjaga-antirez-ds4wa-unsloth-ggufyi-2biteu-yangjahwa-bigyoreul-yoceong/korshunov.aiko2026-07-15T23:30:55.274931+00:00사용자가 Antirez DS4와 Unsloth GGUF의 2비트 양자화 비교를 요청 https://korshunov.ai/hi/article/12217-upyogkrtaa-antirez-ds4-aur-unsloth-gguf-ke-biic-2-bit-kvaanttiikrnn-pr-tulnaa/korshunov.aihi2026-07-15T23:30:55.274931+00:00उपयोगकर्ता Antirez DS4 और Unsloth GGUF के बीच 2-bit क्वांटीकरण पर तुलना का अनुरोध करता है https://korshunov.ai/zh-hans/article/12217-yong-hu-qing-qiu-bi-jiao-antirez-ds4-he-unsloth-gguf-zai-2-bit-liang-hua-xia-de/korshunov.aizh-cn2026-07-15T23:30:55.274931+00:00用户请求比较 Antirez DS4 和 Unsloth GGUF 在 2-bit 量化下的表现 https://korshunov.ai/es/article/12217-el-usuario-solicita-una-comparacion-entre-antirez-ds4-y-unsloth-gguf-con-de-2/korshunov.aies2026-07-15T23:30:55.274931+00:00El usuario solicita una comparación entre Antirez DS4 y Unsloth GGUF con cuantización de 2 bits https://korshunov.ai/ru/article/12217-polzovatel-zaprashivaet-sravnenie-antirez-ds4-i-unsloth-gguf-pri-2-bitnom/korshunov.airu2026-07-15T23:30:55.274931+00:00Пользователь запрашивает сравнение Antirez DS4 и Unsloth GGUF при 2-битном квантовании https://korshunov.ai/ja/article/12217-yu-sa-kaantirez-ds4tounsloth-ggufno2hitsutoliang-zi-hua-tenobi-jiao-worikuesuto/korshunov.aija2026-07-15T23:30:55.274931+00:00ユーザーがAntirez DS4とUnsloth GGUFの2ビット量子化での比較をリクエスト https://korshunov.ai/en/article/12217-user-asks-for-comparison-between-antirez-ds4-and-unsloth-gguf-at-2-bit/korshunov.aien2026-07-15T23:30:55.274931+00:00User asks for comparison between Antirez DS4 and Unsloth GGUF at 2-bit quantization https://korshunov.ai/ar/article/12217-lmstkhdm-ytlb-mqrn-byn-antirez-ds4-w-unsloth-gguf-nd-ltkmym-thnyy-lbt/korshunov.aiar2026-07-15T23:30:55.274931+00:00المستخدم يطلب مقارنة بين Antirez DS4 و Unsloth GGUF عند التكميم ثنائي البت https://korshunov.ai/pt-br/article/12217-usuario-pede-comparacao-entre-antirez-ds4-e-unsloth-gguf-em-quantizacao-de-2/korshunov.aipt2026-07-15T23:30:55.274931+00:00Usuário pede comparação entre Antirez DS4 e Unsloth GGUF em quantização de 2 bits https://korshunov.ai/id/article/12217-pengguna-meminta-perbandingan-antara-antirez-ds4-dan-unsloth-gguf-pada-2-bit/korshunov.aiid2026-07-15T23:30:55.274931+00:00Pengguna meminta perbandingan antara Antirez DS4 dan Unsloth GGUF pada kuantisasi 2-bit https://korshunov.ai/ko/article/12218-tencent-hy3reul-pohamhan-sae-modeli-dyueol-dgx-sparke-jeoghab/korshunov.aiko2026-07-15T23:22:50.760305+00:00Tencent Hy3를 포함한 새 모델이 듀얼 DGX Spark에 적합 https://korshunov.ai/fr/article/12218-de-nouveaux-modeles-dont-tencent-hy3-tiennent-sur-un-dgx-spark-double/korshunov.aifr2026-07-15T23:22:50.760305+00:00De nouveaux modèles, dont Tencent Hy3, tiennent sur un DGX Spark double https://korshunov.ai/ar/article/12218-nmdhj-jdyd-bm-fy-dhlk-tencent-hy3-tts-l-dgx-spark-mzdwj/korshunov.aiar2026-07-15T23:22:50.760305+00:00نماذج جديدة بما في ذلك Tencent Hy3 تتسع لـ DGX Spark مزدوج https://korshunov.ai/hi/article/12218-tencent-hy3-shit-ne-monddl-ddual-dgx-spark-pr-phitt-baitthte-hain/korshunov.aihi2026-07-15T23:22:50.760305+00:00Tencent Hy3 सहित नए मॉडल डुअल DGX Spark पर फिट बैठते हैं https://korshunov.ai/ja/article/12218-tencent-hy3wohan-muxin-moterukateyuarudgx-sparknishou-maru/korshunov.aija2026-07-15T23:22:50.760305+00:00Tencent Hy3を含む新モデルがデュアルDGX Sparkに収まる https://korshunov.ai/es/article/12218-nuevos-modelos-incluido-tencent-hy3-caben-en-dos-dgx-spark/korshunov.aies2026-07-15T23:22:50.760305+00:00Nuevos modelos, incluido Tencent Hy3, caben en dos DGX Spark https://korshunov.ai/id/article/12218-model-baru-termasuk-tencent-hy3-muat-di-dgx-spark-ganda/korshunov.aiid2026-07-15T23:22:50.760305+00:00Model baru termasuk Tencent Hy3 muat di DGX Spark ganda https://korshunov.ai/en/article/12218-new-models-including-tencent-hy3-fit-on-dual-dgx-spark/korshunov.aien2026-07-15T23:22:50.760305+00:00New models including Tencent Hy3 fit on dual DGX Spark https://korshunov.ai/zh-hans/article/12218-bao-gua-teng-xun-hy3-zai-nei-de-xin-mo-xing-ke-gua-pei-shuang-dgx-spark/korshunov.aizh-cn2026-07-15T23:22:50.760305+00:00包括腾讯 Hy3 在内的新模型可适配双 DGX Spark https://korshunov.ai/pt-br/article/12218-novos-modelos-incluindo-tencent-hy3-cabem-em-dois-dgx-spark/korshunov.aipt2026-07-15T23:22:50.760305+00:00Novos modelos, incluindo Tencent Hy3, cabem em dois DGX Spark https://korshunov.ai/ru/article/12218-novye-modeli-vkliuchaia-tencent-hy3-pomeshchaiutsia-na-dva-dgx-spark/korshunov.airu2026-07-15T23:22:50.760305+00:00Новые модели, включая Tencent Hy3, помещаются на два DGX Spark https://korshunov.ai/ar/article/12219-qwen3-5-122b-a10b-b-ud-q2-k-xl-yts-l-64-jyjbyt-mn-dhkr-lwswl-l-shwyy-wyhsn-lm-rf/korshunov.aiar2026-07-15T23:15:13.402299+00:00Qwen3.5 122B A10B بـ UD-Q2_K_XL يتسع لـ 64 جيجابايت من ذاكرة الوصول العشوائي ويحسن المعرفة الداخلية https://korshunov.ai/id/article/12219-qwen3-5-122b-a10b-dengan-ud-q2-k-xl-muat-di-ram-64gb-dan-meningkatkan-internal/korshunov.aiid2026-07-15T23:15:13.402299+00:00Qwen3.5 122B A10B dengan UD-Q2_K_XL muat di RAM 64GB dan meningkatkan pengetahuan internal https://korshunov.ai/fr/article/12219-qwen3-5-122b-a10b-quantise-avec-ud-q2-k-xl-tient-dans-64-go-de-ram-et-ameliore/korshunov.aifr2026-07-15T23:15:13.402299+00:00Qwen3.5 122B A10B quantisé avec UD-Q2_K_XL tient dans 64 Go de RAM et améliore les connaissances internes https://korshunov.ai/es/article/12219-qwen3-5-122b-a10b-en-formato-ud-q2-k-xl-cabe-en-64-gb-de-ram-y-mejora-el-interno/korshunov.aies2026-07-15T23:15:13.402299+00:00Qwen3.5 122B A10B en formato UD-Q2_K_XL cabe en 64 GB de RAM y mejora el conocimiento interno https://korshunov.ai/ru/article/12219-qwen3-5-122b-a10b-v-formate-ud-q2-k-xl-pomeshchaetsia-v-64-gb-ozu-i-uluchshaet/korshunov.airu2026-07-15T23:15:13.402299+00:00Qwen3.5 122B A10B в формате UD-Q2_K_XL помещается в 64 ГБ ОЗУ и улучшает внутренние знания https://korshunov.ai/en/article/12219-qwen3-5-122b-a10b-at-ud-q2-k-xl-fits-in-64gb-ram-and-improves-internal-knowledge/korshunov.aien2026-07-15T23:15:13.402299+00:00Qwen3.5 122B A10B at UD-Q2_K_XL fits in 64GB RAM and improves internal knowledge https://korshunov.ai/ja/article/12219-ud-q2-k-xlteliang-zi-hua-saretaqwen3-5-122b-a10bha64gb-ramnishou-mari-nei-bu-zhi/korshunov.aija2026-07-15T23:15:13.402299+00:00UD-Q2_K_XLで量子化されたQwen3.5 122B A10Bは64GB RAMに収まり、内部知識が向上 https://korshunov.ai/hi/article/12219-ud-q2-k-xl-men-qwen3-5-122b-a10b-64gb-ram-men-phitt-hotaa-hai-aur-aantrik-jnyaan/korshunov.aihi2026-07-15T23:15:13.402299+00:00UD-Q2_K_XL में Qwen3.5 122B A10B, 64GB RAM में फिट होता है और आंतरिक ज्ञान में सुधार करता है https://korshunov.ai/pt-br/article/12219-qwen3-5-122b-a10b-no-formato-ud-q2-k-xl-cabe-em-64-gb-de-ram-e-melhora-o-interno/korshunov.aipt2026-07-15T23:15:13.402299+00:00Qwen3.5 122B A10B no formato UD-Q2_K_XL cabe em 64 GB de RAM e melhora o conhecimento interno https://korshunov.ai/zh-hans/article/12219-ud-q2-k-xl-ge-shi-de-qwen3-5-122b-a10b-ke-rong-na-yu-64gb-ram-bing-ti-sheng-nei/korshunov.aizh-cn2026-07-15T23:15:13.402299+00:00UD-Q2_K_XL 格式的 Qwen3.5 122B A10B 可容纳于 64GB RAM 并提升内部知识 https://korshunov.ai/ko/article/12219-ud-q2-k-xlro-yangjahwadoen-qwen3-5-122b-a10bneun-64gb-rame-majeumyeo-naebu/korshunov.aiko2026-07-15T23:15:13.402299+00:00UD-Q2_K_XL로 양자화된 Qwen3.5 122B A10B는 64GB RAM에 맞으며 내부 지식을 개선합니다 https://korshunov.ai/ko/article/12182-pluralis-research-sobijayong-mac-14dae-wieseo-rl-sahu-hagseub-silhaeng/korshunov.aiko2026-07-15T23:05:29.323742+00:00Pluralis Research, 소비자용 Mac 14대 위에서 RL 사후 학습 실행 https://korshunov.ai/fr/article/12182-pluralis-research-execute-un-post-training-rl-sur-14-mac-grand-public/korshunov.aifr2026-07-15T23:05:29.323742+00:00Pluralis Research exécute un post-training RL sur 14 Mac grand public https://korshunov.ai/hi/article/12182-pluralis-research-ne-14-upbhoktaa-macs-pr-rl-postt-ttrening-claaii/korshunov.aihi2026-07-15T23:05:29.323742+00:00Pluralis Research ने 14 उपभोक्ता Macs पर RL पोस्ट-ट्रेनिंग चलाई https://korshunov.ai/ja/article/12182-pluralis-researchka14tai-noxiao-fei-zhe-xiang-kemacterlhosutotore-ninkuwoshi/korshunov.aija2026-07-15T23:05:29.323742+00:00Pluralis Researchが14台の消費者向けMacでRLポストトレーニングを実行 https://korshunov.ai/id/article/12182-pluralis-research-menjalankan-rl-post-training-pada-14-mac-konsumen/korshunov.aiid2026-07-15T23:05:29.323742+00:00Pluralis Research menjalankan RL post-training pada 14 Mac konsumen https://korshunov.ai/ar/article/12182-pluralis-research-tshgwl-tdryban-lhqan-blt-zyz-lm-zz-l-14-jhz-mac-llmsthlkyn/korshunov.aiar2026-07-15T23:05:29.323742+00:00Pluralis Research تشغّل تدريبًا لاحقًا بالتعزيز المعزز على 14 جهاز Mac للمستهلكين https://korshunov.ai/zh-hans/article/12182-pluralis-research-zai-14-tai-xiao-fei-ji-mac-shang-yun-xing-rl-hou-xun-lian/korshunov.aizh-cn2026-07-15T23:05:29.323742+00:00Pluralis Research 在 14 台消费级 Mac 上运行 RL 后训练 https://korshunov.ai/es/article/12182-pluralis-research-ejecuta-rl-post-entrenamiento-en-14-mac-de-consumo/korshunov.aies2026-07-15T23:05:29.323742+00:00Pluralis Research ejecuta RL post-entrenamiento en 14 Mac de consumo https://korshunov.ai/pt-br/article/12182-pluralis-research-executa-rl-pos-treinamento-em-14-macs-de-consumo/korshunov.aipt2026-07-15T23:05:29.323742+00:00Pluralis Research executa RL pós-treinamento em 14 Macs de consumo https://korshunov.ai/en/article/12182-pluralis-research-runs-rl-post-training-on-14-consumer-macs/korshunov.aien2026-07-15T23:05:29.323742+00:00Pluralis Research runs RL post-training on 14 consumer Macs https://korshunov.ai/ru/article/12182-pluralis-research-provela-rl-postobuchenie-na-14-potrebitelskikh-mac/korshunov.airu2026-07-15T23:05:29.323742+00:00Pluralis Research провела RL постобучение на 14 потребительских Mac https://korshunov.ai/pt-br/article/12188-internscience-lanca-agents-a1-4b-um-modelo-de-4-bilhoes-de-parametros-para-de/korshunov.aipt2026-07-15T22:53:42.522555+00:00InternScience lança Agents-A1-4B, um modelo de 4 bilhões de parâmetros para tarefas agênticas de longo prazo https://korshunov.ai/ar/article/12188-internscience-ttlq-nmwdhj-agents-a1-4b-b-4-mlyrt-m-lm-lmhm-lwkl-twyl-lmd/korshunov.aiar2026-07-15T22:53:42.522555+00:00InternScience تطلق نموذج Agents-A1-4B بـ 4 مليارات معلمة لمهام الوكلاء طويلة المدى https://korshunov.ai/zh-hans/article/12188-internsciencefa-bu-agents-a1-4b-yi-kuan-yong-yu-chang-zhou-qi-zhi-neng-ti-ren-wu/korshunov.aizh-cn2026-07-15T22:53:42.522555+00:00InternScience发布Agents-A1-4B，一款用于长周期智能体任务的40亿参数模型 https://korshunov.ai/hi/article/12188-internscience-ne-agents-a1-4b-jaarii-kiyaa-lnbe-smy-tk-clne-vaale-ejenttik-ke-4b/korshunov.aihi2026-07-15T22:53:42.522555+00:00InternScience ने Agents-A1-4B जारी किया, लंबे समय तक चलने वाले एजेंटिक कार्यों के लिए 4B-पैरामीटर मॉडल https://korshunov.ai/en/article/12188-internscience-releases-agents-a1-4b-a-4b-parameter-model-for-long-horizon-tasks/korshunov.aien2026-07-15T22:53:42.522555+00:00InternScience releases Agents-A1-4B, a 4B-parameter model for long-horizon agentic tasks https://korshunov.ai/ru/article/12188-internscience-vypustila-agents-a1-4b-model-na-4-mlrd-parametrov-dlia-agentnykh/korshunov.airu2026-07-15T22:53:42.522555+00:00InternScience выпустила Agents-A1-4B, модель на 4 млрд параметров для долгосрочных агентных задач https://korshunov.ai/ja/article/12188-internsciencekachang-qi-horaisonnoe-shientotasukuxiang-ke4bharame-tamoteruagents/korshunov.aija2026-07-15T22:53:42.522555+00:00InternScienceが長期ホライゾンのエージェントタスク向け4BパラメータモデルAgents-A1-4Bをリリース https://korshunov.ai/es/article/12188-internscience-lanza-agents-a1-4b-un-modelo-de-4-mil-millones-de-parametros-para/korshunov.aies2026-07-15T22:53:42.522555+00:00InternScience lanza Agents-A1-4B, un modelo de 4 mil millones de parámetros para tareas agénticas de largo plazo https://korshunov.ai/ko/article/12188-internscience-janggi-horaijeun-eijeonteu-jageobyong-4b-paramiteo-model-agents-a1/korshunov.aiko2026-07-15T22:53:42.522555+00:00InternScience, 장기 호라이즌 에이전트 작업용 4B 파라미터 모델 Agents-A1-4B 출시 https://korshunov.ai/fr/article/12188-internscience-lance-agents-a1-4b-un-modele-de-4-milliards-de-parametres-pour-les/korshunov.aifr2026-07-15T22:53:42.522555+00:00InternScience lance Agents-A1-4B, un modèle de 4 milliards de paramètres pour les tâches agentic à long terme https://korshunov.ai/id/article/12188-internscience-merilis-agents-a1-4b-model-4-miliar-parameter-untuk-tugas-agentic/korshunov.aiid2026-07-15T22:53:42.522555+00:00InternScience merilis Agents-A1-4B, model 4 miliar parameter untuk tugas agentic jangka panjang https://korshunov.ai/ar/article/12212-a-l-f-r-e-d-tukthwf-m-rf-lnmdhj-lkbyr-fy-nmdhj-bhjm-2b-ltnfydh-sr/korshunov.aiar2026-07-15T22:44:37.967266+00:00A.L.F.R.E.D. تُكثّف معرفة النماذج الكبيرة في نماذج بحجم 2B لتنفيذ أسرع https://korshunov.ai/id/article/12212-a-l-f-r-e-d-memadatkan-pengetahuan-model-besar-ke-dalam-model-2b-untuk-eksekusi/korshunov.aiid2026-07-15T22:44:37.967266+00:00A.L.F.R.E.D. memadatkan pengetahuan model besar ke dalam model 2B untuk eksekusi lebih cepat https://korshunov.ai/ja/article/12212-a-l-f-r-e-d-ka2bmoterunida-gui-mo-moterunozhi-shi-wozheng-liu-shi-shi-xing-su-du/korshunov.aija2026-07-15T22:44:37.967266+00:00A.L.F.R.E.D.が2Bモデルに大規模モデルの知識を蒸留し、実行速度を向上 https://korshunov.ai/fr/article/12212-a-l-f-r-e-d-condense-le-savoir-des-grands-modeles-en-modeles-de-2b-pour-une-plus/korshunov.aifr2026-07-15T22:44:37.967266+00:00A.L.F.R.E.D. condense le savoir des grands modèles en modèles de 2B pour une exécution plus rapide https://korshunov.ai/es/article/12212-a-l-f-r-e-d-destila-el-conocimiento-de-modelos-grandes-en-modelos-de-2b-para-una/korshunov.aies2026-07-15T22:44:37.967266+00:00A.L.F.R.E.D. destila el conocimiento de modelos grandes en modelos de 2B para una ejecución más rápida https://korshunov.ai/pt-br/article/12212-a-l-f-r-e-d-destila-conhecimento-de-modelos-grandes-em-modelos-de-2b-para-mais/korshunov.aipt2026-07-15T22:44:37.967266+00:00A.L.F.R.E.D. destila conhecimento de modelos grandes em modelos de 2B para execução mais rápida https://korshunov.ai/en/article/12212-a-l-f-r-e-d-distills-large-model-knowledge-into-2b-models-for-faster-execution/korshunov.aien2026-07-15T22:44:37.967266+00:00A.L.F.R.E.D. distills large model knowledge into 2B models for faster execution https://korshunov.ai/ru/article/12212-a-l-f-r-e-d-distilliruet-znaniia-bolshikh-modelei-v-modeli-na-2b-dlia/korshunov.airu2026-07-15T22:44:37.967266+00:00A.L.F.R.E.D. дистиллирует знания больших моделей в модели на 2B для ускоренного выполнения https://korshunov.ai/ko/article/12212-a-l-f-r-e-d-daegyumo-model-jisigeul-2b-modele-jeungryuhayeo-silhaeng-sogdo/korshunov.aiko2026-07-15T22:44:37.967266+00:00A.L.F.R.E.D., 대규모 모델 지식을 2B 모델에 증류하여 실행 속도 향상 https://korshunov.ai/zh-hans/article/12212-a-l-f-r-e-d-jiang-da-xing-mo-xing-zhi-shi-zheng-liu-zhi-2bmo-xing-yi-shi-xian/korshunov.aizh-cn2026-07-15T22:44:37.967266+00:00A.L.F.R.E.D.将大型模型知识蒸馏至2B模型以实现更快执行 https://korshunov.ai/hi/article/12212-a-l-f-r-e-d-bdde-monddl-ke-jnyaan-ko-tej-nisspaadn-ke-lie-2b-monddlon-men-krtaa/korshunov.aihi2026-07-15T22:44:37.967266+00:00A.L.F.R.E.D. बड़े मॉडल के ज्ञान को तेज़ निष्पादन के लिए 2B मॉडलों में संक्षिप्त करता है https://korshunov.ai/ko/article/12211-singking-meosini-coecoyi-opeunsoseu-model-inkeulringe-daehan-bencimakeu-gonggae/korshunov.aiko2026-07-15T22:37:01.162206+00:00싱킹 머신이 최초의 오픈소스 모델 인클링에 대한 벤치마크 공개 https://korshunov.ai/ja/article/12211-shinkinkumashinkachu-noo-hunso-sumoteru-inkurinku-nohenchima-kuwogong-kai/korshunov.aija2026-07-15T22:37:01.162206+00:00シンキングマシンが初のオープンソースモデル「インクリング」のベンチマークを公開 https://korshunov.ai/en/article/12211-thinking-machine-releases-benchmarks-for-its-first-open-source-model-inkling/korshunov.aien2026-07-15T22:37:01.162206+00:00Thinking Machine releases benchmarks for its first open-source model Inkling https://korshunov.ai/ru/article/12211-thinking-machine-publikuet-benchmarki-dlia-svoei-pervoi-otkrytoi-modeli-inkling/korshunov.airu2026-07-15T22:37:01.162206+00:00Thinking Machine публикует бенчмарки для своей первой открытой модели Inkling https://korshunov.ai/es/article/12211-thinking-machine-publica-benchmarks-para-su-primer-modelo-de-codigo-abierto/korshunov.aies2026-07-15T22:37:01.162206+00:00Thinking Machine publica benchmarks para su primer modelo de código abierto Inkling https://korshunov.ai/zh-hans/article/12211-thinking-machine-fa-bu-qi-shou-ge-kai-yuan-mo-xing-inkling-de-ji-zhun-ce-shi/korshunov.aizh-cn2026-07-15T22:37:01.162206+00:00Thinking Machine 发布其首个开源模型 Inkling 的基准测试 https://korshunov.ai/pt-br/article/12211-thinking-machine-publica-benchmarks-para-seu-primeiro-modelo-de-codigo-aberto/korshunov.aipt2026-07-15T22:37:01.162206+00:00Thinking Machine publica benchmarks para seu primeiro modelo de código aberto, Inkling https://korshunov.ai/hi/article/12211-thinking-mshiin-ne-apne-phle-opn-sors-monddl-inkling-ke-lie-bencmaarks-jaarii/korshunov.aihi2026-07-15T22:37:01.162206+00:00थिंकिंग मशीन ने अपने पहले ओपन-सोर्स मॉडल इंकलिंग के लिए बेंचमार्क्स जारी किए https://korshunov.ai/fr/article/12211-thinking-machine-publie-les-benchmarks-de-son-premier-modele-open-source-inkling/korshunov.aifr2026-07-15T22:37:01.162206+00:00Thinking Machine publie les benchmarks de son premier modèle open-source Inkling https://korshunov.ai/id/article/12211-thinking-machine-merilis-benchmark-untuk-model-open-source-pertamanya-inkling/korshunov.aiid2026-07-15T22:37:01.162206+00:00Thinking Machine merilis benchmark untuk model open-source pertamanya Inkling https://korshunov.ai/ar/article/12211-tyknj-mshyn-ttlq-m-yyr-ld-lnmwdhjh-lwl-mftwh-lmsdr-nklng/korshunov.aiar2026-07-15T22:37:01.162206+00:00تيكنج ماشين تطلق معايير الأداء لنموذجها الأول مفتوح المصدر إنكلنغ https://korshunov.ai/es/article/12215-inkling-de-thinking-machines-ocupa-el-puesto-5-entre-todos-los-modelos-de-peso/korshunov.aies2026-07-15T22:32:46.991992+00:00Inkling de Thinking Machines ocupa el puesto #5 entre todos los modelos de peso abierto https://korshunov.ai/zh-hans/article/12215-thinking-machines-de-inkling-zai-suo-you-kai-yuan-quan-zhong-mo-xing-zhong-pai-5/korshunov.aizh-cn2026-07-15T22:32:46.991992+00:00Thinking Machines 的 Inkling 在所有开源权重模型中排名第 5 https://korshunov.ai/id/article/12215-inkling-oleh-thinking-machines-peringkat-5-di-antara-semua-model-bobot-terbuka/korshunov.aiid2026-07-15T22:32:46.991992+00:00Inkling oleh Thinking Machines peringkat #5 di antara semua model bobot terbuka https://korshunov.ai/ar/article/12215-nklng-mn-thynkyng-mshynz-thtl-lmrtb-lkhms-byn-jmy-lnmdhj-dht-lwzn-lmftwh/korshunov.aiar2026-07-15T22:32:46.991992+00:00إنكلنغ من ثينكينغ ماشينز تحتل المرتبة الخامسة بين جميع النماذج ذات الأوزان المفتوحة https://korshunov.ai/fr/article/12215-inkling-de-thinking-machines-se-classe-ndeg5-parmi-tous-les-modeles-a-poids/korshunov.aifr2026-07-15T22:32:46.991992+00:00Inkling de Thinking Machines se classe n°5 parmi tous les modèles à poids ouverts https://korshunov.ai/ko/article/12215-thinking-machinesyi-inklingi-modeun-opeun-gajungci-model-jung-5wie-raengkeu/korshunov.aiko2026-07-15T22:32:46.991992+00:00Thinking Machines의 Inkling이 모든 오픈 가중치 모델 중 5위에 랭크 https://korshunov.ai/en/article/12215-inkling-by-thinking-machines-ranks-5-among-all-open-weight-models/korshunov.aien2026-07-15T22:32:46.991992+00:00Inkling by Thinking Machines ranks #5 among all open weight models https://korshunov.ai/ja/article/12215-thinking-machinesnoinklingkasuhetenoo-hunueitomoterunozhong-te5wei-nirankuin/korshunov.aija2026-07-15T22:32:46.991992+00:00Thinking MachinesのInklingがすべてのオープンウェイトモデルの中で5位にランクイン https://korshunov.ai/pt-br/article/12215-inkling-da-thinking-machines-ocupa-o-5-entre-todos-os-modelos-de-peso-aberto/korshunov.aipt2026-07-15T22:32:46.991992+00:00Inkling da Thinking Machines ocupa o #5 entre todos os modelos de peso aberto https://korshunov.ai/ru/article/12215-inkling-ot-thinking-machines-zanimaet-5-e-mesto-sredi-vsekh-modelei-s-otkrytym/korshunov.airu2026-07-15T22:32:46.991992+00:00Inkling от Thinking Machines занимает 5-е место среди всех моделей с открытым весом https://korshunov.ai/hi/article/12215-sbhii-opn-vett-monddl-men-thinking-mshiins-kaa-inkling-5-pr-raink-kiyaa-gyaa/korshunov.aihi2026-07-15T22:32:46.991992+00:00सभी ओपन वेट मॉडल में थिंकिंग मशीन्स का इंकलिंग #5 पर रैंक किया गया https://korshunov.ai/fr/article/12216-xai-open-source-grok-build-sous-licence-apache-2-0/korshunov.aifr2026-07-15T22:27:00.105788+00:00xAI open source Grok Build sous licence Apache 2.0 https://korshunov.ai/ko/article/12216-xai-grok-buildreul-apache-2-0-raiseonseuro-opeunsoseuhwa/korshunov.aiko2026-07-15T22:27:00.105788+00:00xAI, Grok Build를 Apache 2.0 라이선스로 오픈소스화 https://korshunov.ai/ru/article/12216-xai-otkryla-iskhodnyi-kod-grok-build-pod-litsenziei-apache-2-0/korshunov.airu2026-07-15T22:27:00.105788+00:00xAI открыла исходный код Grok Build под лицензией Apache 2.0 https://korshunov.ai/es/article/12216-xai-abre-el-codigo-fuente-de-grok-build-bajo-la-licencia-apache-2-0/korshunov.aies2026-07-15T22:27:00.105788+00:00xAI abre el código fuente de Grok Build bajo la licencia Apache 2.0 https://korshunov.ai/pt-br/article/12216-xai-abre-o-codigo-fonte-do-grok-build-sob-a-licenca-apache-2-0/korshunov.aipt2026-07-15T22:27:00.105788+00:00xAI abre o código-fonte do Grok Build sob a licença Apache 2.0 https://korshunov.ai/zh-hans/article/12216-xai-zai-apache-2-0-xu-ke-xia-kai-yuan-grok-build/korshunov.aizh-cn2026-07-15T22:27:00.105788+00:00xAI 在 Apache 2.0 许可下开源 Grok Build https://korshunov.ai/en/article/12216-xai-open-sources-grok-build-under-apache-2-0-license/korshunov.aien2026-07-15T22:27:00.105788+00:00xAI open sources Grok Build under Apache 2.0 license https://korshunov.ai/hi/article/12216-xai-ne-apache-2-0-laaisens-ke-tht-grok-build-kaa-srot-kodd-opn-sors-kiyaa/korshunov.aihi2026-07-15T22:27:00.105788+00:00xAI ने Apache 2.0 लाइसेंस के तहत Grok Build का स्रोत कोड ओपन सोर्स किया https://korshunov.ai/ar/article/12216-xai-tfth-msdr-grok-build-tht-rkhs-apache-2-0/korshunov.aiar2026-07-15T22:27:00.105788+00:00xAI تفتح مصدر Grok Build تحت رخصة Apache 2.0 https://korshunov.ai/ja/article/12216-xaikagrok-buildwoapache-2-0raisensunoxia-teo-hunso-suhua/korshunov.aija2026-07-15T22:27:00.105788+00:00xAIがGrok BuildをApache 2.0ライセンスの下でオープンソース化 https://korshunov.ai/id/article/12216-xai-membuka-sumber-grok-build-di-bawah-lisensi-apache-2-0/korshunov.aiid2026-07-15T22:27:00.105788+00:00xAI membuka sumber Grok Build di bawah lisensi Apache 2.0 https://korshunov.ai/ko/article/12214-llama-cpp-b10032e-cuda-ggml-op-lightning-indexer-cuga/korshunov.aiko2026-07-15T22:21:48.347098+00:00llama.cpp b10032에 CUDA GGML_OP_LIGHTNING_INDEXER 추가 https://korshunov.ai/fr/article/12214-llama-cpp-b10032-ajoute-ggml-op-lightning-indexer-cuda/korshunov.aifr2026-07-15T22:21:48.347098+00:00llama.cpp b10032 ajoute GGML_OP_LIGHTNING_INDEXER CUDA https://korshunov.ai/id/article/12214-llama-cpp-b10032-menambahkan-ggml-op-lightning-indexer-cuda/korshunov.aiid2026-07-15T22:21:48.347098+00:00llama.cpp b10032 menambahkan GGML_OP_LIGHTNING_INDEXER CUDA https://korshunov.ai/ar/article/12214-ydyf-llama-cpp-b10032-d-m-ggml-op-lightning-indexer-l-cuda/korshunov.aiar2026-07-15T22:21:48.347098+00:00يضيف llama.cpp b10032 دعم GGML_OP_LIGHTNING_INDEXER لـ CUDA https://korshunov.ai/es/article/12214-llama-cpp-b10032-anade-cuda-ggml-op-lightning-indexer/korshunov.aies2026-07-15T22:21:48.347098+00:00llama.cpp b10032 añade CUDA GGML_OP_LIGHTNING_INDEXER https://korshunov.ai/pt-br/article/12214-llama-cpp-b10032-adiciona-cuda-ggml-op-lightning-indexer/korshunov.aipt2026-07-15T22:21:48.347098+00:00llama.cpp b10032 adiciona CUDA GGML_OP_LIGHTNING_INDEXER https://korshunov.ai/zh-hans/article/12214-llama-cpp-b10032-tian-jia-liao-cuda-ggml-op-lightning-indexer/korshunov.aizh-cn2026-07-15T22:21:48.347098+00:00llama.cpp b10032 添加了 CUDA GGML_OP_LIGHTNING_INDEXER https://korshunov.ai/hi/article/12214-llama-cpp-b10032-men-cuda-ggml-op-lightning-indexer-joddaa-gyaa/korshunov.aihi2026-07-15T22:21:48.347098+00:00llama.cpp b10032 में CUDA GGML_OP_LIGHTNING_INDEXER जोड़ा गया https://korshunov.ai/ja/article/12214-llama-cpp-b10032kacuda-ggml-op-lightning-indexerwozhui-jia/korshunov.aija2026-07-15T22:21:48.347098+00:00llama.cpp b10032がCUDA GGML_OP_LIGHTNING_INDEXERを追加 https://korshunov.ai/ru/article/12214-llama-cpp-b10032-dobavliaet-cuda-ggml-op-lightning-indexer/korshunov.airu2026-07-15T22:21:48.347098+00:00llama.cpp b10032 добавляет CUDA GGML_OP_LIGHTNING_INDEXER https://korshunov.ai/en/article/12214-llama-cpp-b10032-adds-cuda-ggml-op-lightning-indexer/korshunov.aien2026-07-15T22:21:48.347098+00:00llama.cpp b10032 adds CUDA GGML_OP_LIGHTNING_INDEXER https://korshunov.ai/en/article/12213-llama-cpp-b10034-excludes-adreno-a7x-from-moe-kernels-to-prevent-weight/korshunov.aien2026-07-15T22:11:13.334315+00:00llama.cpp b10034 excludes Adreno A7x from MoE kernels to prevent weight corruption https://korshunov.ai/es/article/12213-llama-cpp-b10034-excluye-a-adreno-a7x-de-los-nucleos-moe-para-evitar-la-de-pesos/korshunov.aies2026-07-15T22:11:13.334315+00:00llama.cpp b10034 excluye a Adreno A7x de los núcleos MoE para evitar la corrupción de pesos https://korshunov.ai/ru/article/12213-llama-cpp-b10034-iskliuchaet-adreno-a7x-iz-moe-iader-dlia-predotvrashcheniia/korshunov.airu2026-07-15T22:11:13.334315+00:00llama.cpp b10034 исключает Adreno A7x из MoE-ядер для предотвращения повреждения весов https://korshunov.ai/zh-hans/article/12213-llama-cpp-b10034-pai-chu-adreno-a7x-shi-yong-moe-nei-he-yi-fang-zhi-quan-zhong/korshunov.aizh-cn2026-07-15T22:11:13.334315+00:00llama.cpp b10034 排除 Adreno A7x 使用 MoE 内核以防止权重损坏 https://korshunov.ai/hi/article/12213-llama-cpp-b10034-vjn-ke-kssrnn-ko-rokne-ke-lie-moe-kernels-se-adreno-a7x-ko-hai/korshunov.aihi2026-07-15T22:11:13.334315+00:00llama.cpp b10034 वजन के क्षरण को रोकने के लिए MoE kernels से Adreno A7x को बाहर करता है https://korshunov.ai/pt-br/article/12213-llama-cpp-b10034-exclui-adreno-a7x-dos-kernels-moe-para-evitar-corrupcao-de/korshunov.aipt2026-07-15T22:11:13.334315+00:00llama.cpp b10034 exclui Adreno A7x dos kernels MoE para evitar corrupção de pesos https://korshunov.ai/ja/article/12213-llama-cpp-b10034-zhong-mipo-sun-wofang-kutamemoeka-nerukaraadreno-a7xwochu-wai/korshunov.aija2026-07-15T22:11:13.334315+00:00llama.cpp b10034、重み破損を防ぐためMoEカーネルからAdreno A7xを除外 https://korshunov.ai/ko/article/12213-llama-cpp-b10034-gajungci-sonsangeul-bangjihagi-wihae-moe-keoneoleseo-adreno-a7x/korshunov.aiko2026-07-15T22:11:13.334315+00:00llama.cpp b10034, 가중치 손상을 방지하기 위해 MoE 커널에서 Adreno A7x 제외 https://korshunov.ai/ar/article/12213-ystb-d-llama-cpp-b10034-shryh-adreno-a7x-mn-nw-moe-lmn-tlf-lwzn/korshunov.aiar2026-07-15T22:11:13.334315+00:00يستبعد llama.cpp b10034 شريحة Adreno A7x من نوى MoE لمنع تلف الأوزان https://korshunov.ai/id/article/12213-llama-cpp-b10034-mengecualikan-adreno-a7x-dari-kernel-moe-untuk-mencegah-korupsi/korshunov.aiid2026-07-15T22:11:13.334315+00:00llama.cpp b10034 mengecualikan Adreno A7x dari kernel MoE untuk mencegah korupsi bobot https://korshunov.ai/fr/article/12213-llama-cpp-b10034-exclut-adreno-a7x-des-noyaux-moe-pour-eviter-la-corruption-des/korshunov.aifr2026-07-15T22:11:13.334315+00:00llama.cpp b10034 exclut Adreno A7x des noyaux MoE pour éviter la corruption des poids https://korshunov.ai/hi/article/12201-rwkv-ek-rnn-hai-jismen-transformer-kii-trh-smaanaantr-krne-yogytaa-aur-utkrsstt/korshunov.aihi2026-07-15T21:50:24.663370+00:00RWKV एक RNN है जिसमें Transformer की तरह समानांतर करने योग्यता और उत्कृष्ट LLM प्रदर्शन है https://korshunov.ai/ja/article/12201-rwkvhayou-retallmxing-neng-wochi-chi-transformernoyounibing-lie-hua-ke-neng/korshunov.aija2026-07-15T21:50:24.663370+00:00RWKVは優れたLLM性能を持ち、Transformerのように並列化可能なRNN https://korshunov.ai/pt-br/article/12201-rwkv-e-uma-rnn-com-excelente-desempenho-de-llm-e-paralelizavel-como-um/korshunov.aipt2026-07-15T21:50:24.663370+00:00RWKV é uma RNN com excelente desempenho de LLM e paralelizável como um Transformer https://korshunov.ai/zh-hans/article/12201-rwkv-shi-yi-chong-ju-you-transformer-ji-bing-xing-neng-li-qie-da-yu-yan-mo-xing/korshunov.aizh-cn2026-07-15T21:50:24.663370+00:00RWKV 是一种具有 Transformer 级并行能力且大语言模型性能优异的 RNN https://korshunov.ai/en/article/12201-rwkv-is-an-rnn-with-great-llm-performance-and-parallelizable-like-a-transformer/korshunov.aien2026-07-15T21:50:24.663370+00:00RWKV is an RNN with great LLM performance and parallelizable like a Transformer https://korshunov.ai/id/article/12201-rwkv-adalah-rnn-dengan-kinerja-llm-yang-hebat-dan-dapat-diparalelkan-seperti/korshunov.aiid2026-07-15T21:50:24.663370+00:00RWKV adalah RNN dengan kinerja LLM yang hebat dan dapat diparalelkan seperti Transformer https://korshunov.ai/ru/article/12201-rwkv-rnn-s-proizvoditelnostiu-llm-urovnia-transformer-i-vozmozhnostiu/korshunov.airu2026-07-15T21:50:24.663370+00:00RWKV — RNN с производительностью LLM уровня Transformer и возможностью параллелизации https://korshunov.ai/es/article/12201-rwkv-es-una-rnn-con-gran-rendimiento-de-llm-y-paralelizable-como-un-transformer/korshunov.aies2026-07-15T21:50:24.663370+00:00RWKV es una RNN con gran rendimiento de LLM y paralelizable como un Transformer https://korshunov.ai/ko/article/12201-rwkvneun-transformerceoreom-byeongryeolhwa-ganeunghamyeonseodo-ddwieonan-llm-rnn/korshunov.aiko2026-07-15T21:50:24.663370+00:00RWKV는 Transformer처럼 병렬화 가능하면서도 뛰어난 LLM 성능을 가진 RNN https://korshunov.ai/fr/article/12201-rwkv-est-un-rnn-avec-de-grandes-performances-llm-et-parallelisable-comme-un/korshunov.aifr2026-07-15T21:50:24.663370+00:00RWKV est un RNN avec de grandes performances LLM et parallélisable comme un Transformer https://korshunov.ai/ar/article/12201-rwkv-hw-shbk-sby-mtkrr-bd-nmwdhj-lgwy-kbyr-mmtz-wqbl-lltwzy-mthl-lmhwwl/korshunov.aiar2026-07-15T21:50:24.663370+00:00RWKV هو شبكة عصبية متكررة بأداء نموذج لغوي كبير ممتاز وقابل للتوازي مثل المحوّل https://korshunov.ai/ko/article/12197-gaein-yeongujaga-wave-field-llm-culsi-doteu-gob-eotensyeoneul-fft-pasuryeomeuro/korshunov.aiko2026-07-15T21:46:24.333116+00:00개인 연구자가 Wave Field LLM 출시, 도트 곱 어텐션을 FFT 파수렴으로 대체 https://korshunov.ai/fr/article/12197-un-chercheur-independant-lance-wave-field-llm-remplacant-l-attention-par-produit/korshunov.aifr2026-07-15T21:46:24.333116+00:00Un chercheur indépendant lance Wave Field LLM, remplaçant l'attention par produit scalaire par une convolution d'onde FFT https://korshunov.ai/ru/article/12197-issledovatel-odinochka-vypustil-wave-field-llm-zameniv-tochechnoe-proizvedenie/korshunov.airu2026-07-15T21:46:24.333116+00:00Исследователь-одиночка выпустил Wave Field LLM, заменив точечное произведение внимания на FFT-волновую свёртку https://korshunov.ai/en/article/12197-solo-researcher-releases-wave-field-llm-replacing-dot-product-attention-with-fft/korshunov.aien2026-07-15T21:46:24.333116+00:00Solo researcher releases Wave Field LLM, replacing dot-product attention with FFT wave convolution https://korshunov.ai/es/article/12197-un-investigador-independiente-presenta-wave-field-llm-reemplazando-la-atencion/korshunov.aies2026-07-15T21:46:24.333116+00:00Un investigador independiente presenta Wave Field LLM, reemplazando la atención de producto punto por convolución de ondas FFT https://korshunov.ai/zh-hans/article/12197-du-li-yan-jiu-ren-yuan-fa-bu-wave-field-llm-yong-fftbo-juan-ji-qu-dai-dian-ji-yi/korshunov.aizh-cn2026-07-15T21:46:24.333116+00:00独立研究人员发布Wave Field LLM，用FFT波卷积取代点积注意力 https://korshunov.ai/pt-br/article/12197-pesquisador-independente-lanca-wave-field-llm-substituindo-atencao-de-produto-de/korshunov.aipt2026-07-15T21:46:24.333116+00:00Pesquisador independente lança Wave Field LLM, substituindo atenção de produto interno por convolução de onda via FFT https://korshunov.ai/hi/article/12197-ekl-shodhkrtaa-ne-wave-field-llm-jaarii-kiyaa-ddontt-proddktt-ettenshn-ko-fft-se/korshunov.aihi2026-07-15T21:46:24.333116+00:00एकल शोधकर्ता ने Wave Field LLM जारी किया, डॉट-प्रोडक्ट एटेंशन को FFT वेव कन्वोल्यूशन से बदला https://korshunov.ai/ar/article/12197-bhth-mstql-ytlq-wave-field-llm-mstbdlan-ntbh-hsl-ldrb-lnqty-btlfyf-mwj-fft/korshunov.aiar2026-07-15T21:46:24.333116+00:00باحث مستقل يطلق Wave Field LLM، مستبدلاً انتباه حاصل الضرب النقطي بتلافيف موجة FFT https://korshunov.ai/ja/article/12197-ge-ren-yan-jiu-zhe-kawave-field-llmworiri-su-totsutoji-atenshiyonwofftbo-die/korshunov.aija2026-07-15T21:46:24.333116+00:00個人研究者がWave Field LLMをリリース、ドット積アテンションをFFT波畳み込みで置換 https://korshunov.ai/id/article/12197-peneliti-independen-merilis-wave-field-llm-mengganti-perhatian-hasil-titik-fft/korshunov.aiid2026-07-15T21:46:24.333116+00:00Peneliti independen merilis Wave Field LLM, mengganti perhatian hasil titik dengan konvolusi gelombang FFT https://korshunov.ai/id/article/12200-google-memperbarui-template-obrolan-gemma-4-dengan-perbaikan-pemanggilan-alat-4/korshunov.aiid2026-07-15T21:28:49.983129+00:00Google memperbarui template obrolan Gemma 4 dengan perbaikan pemanggilan alat dan dukungan Flash Attention 4 https://korshunov.ai/fr/article/12200-google-met-a-jour-les-modeles-de-chat-de-gemma-4-avec-des-corrections-d-appel-d/korshunov.aifr2026-07-15T21:28:49.983129+00:00Google met à jour les modèles de chat de Gemma 4 avec des corrections d'appel d'outils et le support de Flash Attention 4 https://korshunov.ai/ko/article/12200-google-gemma-4-caeting-tempeulris-eobdeiteu-dogu-hocul-sujeong-mic-flash-4-cuga/korshunov.aiko2026-07-15T21:28:49.983129+00:00Google, Gemma 4 채팅 템플릿 업데이트: 도구 호출 수정 및 Flash Attention 4 지원 추가 https://korshunov.ai/ja/article/12200-googlekagemma-4nochiyatsutotenhure-towogeng-xin-shi-tsu-ruhu-hichu-shinoxiu-jia/korshunov.aija2026-07-15T21:28:49.983129+00:00GoogleがGemma 4のチャットテンプレートを更新し、ツール呼び出しの修正とFlash Attention 4サポートを追加 https://korshunov.ai/hi/article/12200-google-ne-ttuul-konling-phiks-aur-flash-attention-4-sportt-ke-saath-gemma-4-ko/korshunov.aihi2026-07-15T21:28:49.983129+00:00Google ने टूल कॉलिंग फिक्स और Flash Attention 4 सपोर्ट के साथ Gemma 4 चैट टेम्पलेट्स को अपडेट किया https://korshunov.ai/ar/article/12200-jwjl-thdth-qwlb-ldrdsh-l-gemma-4-bslht-std-ldwt-wd-m-flash-attention-4/korshunov.aiar2026-07-15T21:28:49.983129+00:00جوجل تحدث قوالب الدردشة لـ Gemma 4 بإصلاحات استدعاء الأدوات ودعم Flash Attention 4 https://korshunov.ai/pt-br/article/12200-google-atualiza-os-modelos-de-chat-do-gemma-4-com-correcoes-de-chamada-de-e-ao-4/korshunov.aipt2026-07-15T21:28:49.983129+00:00Google atualiza os modelos de chat do Gemma 4 com correções de chamada de ferramentas e suporte ao Flash Attention 4 https://korshunov.ai/es/article/12200-google-actualiza-las-plantillas-de-chat-de-gemma-4-con-correcciones-de-llamada-a/korshunov.aies2026-07-15T21:28:49.983129+00:00Google actualiza las plantillas de chat de Gemma 4 con correcciones de llamada a herramientas y soporte para Flash Attention 4 https://korshunov.ai/ru/article/12200-google-obnovliaet-shablony-chata-gemma-4-s-ispravleniiami-vyzova-instrumentov-i/korshunov.airu2026-07-15T21:28:49.983129+00:00Google обновляет шаблоны чата Gemma 4 с исправлениями вызова инструментов и поддержкой Flash Attention 4 https://korshunov.ai/en/article/12200-google-updates-gemma-4-chat-templates-with-tool-calling-fixes-and-flash-4/korshunov.aien2026-07-15T21:28:49.983129+00:00Google updates Gemma 4 chat templates with tool calling fixes and Flash Attention 4 support https://korshunov.ai/zh-hans/article/12200-google-geng-xin-gemma-4-liao-tian-mo-ban-xiu-fu-gong-ju-diao-yong-wen-ti-bing-4/korshunov.aizh-cn2026-07-15T21:28:49.983129+00:00Google 更新 Gemma 4 聊天模板，修复工具调用问题并支持 Flash Attention 4 https://korshunov.ai/ja/article/12206-yan-jiu-zhe-kadnazhe-rizhi-woshi-tsutemi-mi-metsuse-shiwonanomo-rusufu-hao-to/korshunov.aija2026-07-15T21:22:49.914855+00:00研究者がDNA折り紙を使って秘密メッセージをナノモールス符号でエンコード https://korshunov.ai/ru/article/12206-issledovateli-ispolzuiut-dnk-origami-dlia-kodirovaniia-sekretnykh-soobshchenii-v/korshunov.airu2026-07-15T21:22:49.914855+00:00Исследователи используют ДНК-оригами для кодирования секретных сообщений в наноазбуку Морзе https://korshunov.ai/hi/article/12206-shodhkrtaa-ddiiene-origaamii-kaa-upyog-krke-gopniiy-sndeshon-ko-naino-monrs-kodd/korshunov.aihi2026-07-15T21:22:49.914855+00:00शोधकर्ता डीएनए ओरिगामी का उपयोग करके गोपनीय संदेशों को नैनो-मॉर्स कोड के रूप में एन्कोड कर रहे हैं https://korshunov.ai/es/article/12206-investigadores-usan-origami-de-adn-para-codificar-mensajes-secretos-como-codigo/korshunov.aies2026-07-15T21:22:49.914855+00:00Investigadores usan origami de ADN para codificar mensajes secretos como código Morse a nanoescala https://korshunov.ai/pt-br/article/12206-pesquisadores-usam-origami-de-dna-para-codificar-mensagens-secretas-como-codigo/korshunov.aipt2026-07-15T21:22:49.914855+00:00Pesquisadores usam origami de DNA para codificar mensagens secretas como código Morse em nanoescala https://korshunov.ai/zh-hans/article/12206-yan-jiu-ren-yuan-li-yong-dnazhe-zhi-jiang-mi-mi-xin-xi-bian-ma-wei-na-mi-mo-er/korshunov.aizh-cn2026-07-15T21:22:49.914855+00:00研究人员利用DNA折纸将秘密信息编码为纳米莫尔斯电码 https://korshunov.ai/en/article/12206-researchers-use-dna-origami-to-encode-secret-messages-as-nano-morse-code/korshunov.aien2026-07-15T21:22:49.914855+00:00Researchers use DNA origami to encode secret messages as nano-Morse code https://korshunov.ai/ko/article/12206-yeongujadeuli-dna-origamiro-bimil-mesijireul-nano-moseu-buhoro-inkoding/korshunov.aiko2026-07-15T21:22:49.914855+00:00연구자들이 DNA 오리가미로 비밀 메시지를 나노 모스 부호로 인코딩 https://korshunov.ai/fr/article/12206-les-chercheurs-utilisent-l-origami-en-adn-pour-encoder-des-messages-secrets-en/korshunov.aifr2026-07-15T21:22:49.914855+00:00Les chercheurs utilisent l'origami en ADN pour encoder des messages secrets en code Morse nano https://korshunov.ai/id/article/12206-peneliti-menggunakan-origami-dna-untuk-mengkodekan-pesan-rahasia-sebagai-kode/korshunov.aiid2026-07-15T21:22:49.914855+00:00Peneliti menggunakan origami DNA untuk mengkodekan pesan rahasia sebagai kode Morse nano https://korshunov.ai/ar/article/12206-lbhthwn-ystkhdmwn-wrygmy-lhmd-lnwwy-ltshfyr-lrsyl-lsry-krmz-mwrs-lnnwy/korshunov.aiar2026-07-15T21:22:49.914855+00:00الباحثون يستخدمون أوريغامي الحمض النووي لتشفير الرسائل السرية كرمز مورس النانوي https://korshunov.ai/ar/article/12205-nqsh-hwl-bn-khbyr-dhk-stn-y-mhly-lmnst-brmjy-m-qd/korshunov.aiar2026-07-15T21:17:55.596351+00:00نقاش حول بناء خبير ذكاء اصطناعي محلي لمنصات برمجية معقدة https://korshunov.ai/id/article/12205-diskusi-tentang-membangun-ahli-ai-lokal-untuk-platform-perangkat-lunak-kompleks/korshunov.aiid2026-07-15T21:17:55.596351+00:00Diskusi tentang membangun ahli AI lokal untuk platform perangkat lunak kompleks https://korshunov.ai/fr/article/12205-discussion-sur-la-creation-d-un-expert-ia-local-pour-des-plateformes-logicielles/korshunov.aifr2026-07-15T21:17:55.596351+00:00Discussion sur la création d'un expert IA local pour des plateformes logicielles complexes https://korshunov.ai/ko/article/12205-bogjabhan-sopeuteuweeo-peulraespomeul-wihan-rokeol-ai-jeonmunga-gucuge-daehan/korshunov.aiko2026-07-15T21:17:55.596351+00:00복잡한 소프트웨어 플랫폼을 위한 로컬 AI 전문가 구축에 대한 논의 https://korshunov.ai/ja/article/12205-fu-za-nasohutoueahuratsutohuo-muxiang-kenoro-karuaizhuan-men-jia-gou-zhu-niguan/korshunov.aija2026-07-15T21:17:55.596351+00:00複雑なソフトウェアプラットフォーム向けのローカルAI専門家構築に関する議論 https://korshunov.ai/hi/article/12205-jttil-sonphttveyr-plettphonrm-ke-lie-sthaaniiy-ai-vishessjny-bnaane-pr-crcaa/korshunov.aihi2026-07-15T21:17:55.596351+00:00जटिल सॉफ़्टवेयर प्लेटफ़ॉर्म के लिए स्थानीय AI विशेषज्ञ बनाने पर चर्चा https://korshunov.ai/pt-br/article/12205-discussao-sobre-a-construcao-de-um-especialista-em-ia-local-para-plataformas-de/korshunov.aipt2026-07-15T21:17:55.596351+00:00Discussão sobre a construção de um especialista em IA local para plataformas de software complexas https://korshunov.ai/zh-hans/article/12205-guan-yu-wei-fu-za-ruan-jian-ping-tai-gou-jian-ben-di-aizhuan-jia-de-tao-lun/korshunov.aizh-cn2026-07-15T21:17:55.596351+00:00关于为复杂软件平台构建本地AI专家的讨论 https://korshunov.ai/ru/article/12205-obsuzhdenie-sozdaniia-lokalnogo-ii-eksperta-dlia-slozhnykh-programmnykh-platform/korshunov.airu2026-07-15T21:17:55.596351+00:00Обсуждение создания локального ИИ-эксперта для сложных программных платформ https://korshunov.ai/en/article/12205-discussion-on-building-a-local-ai-expert-for-complex-software-platforms/korshunov.aien2026-07-15T21:17:55.596351+00:00Discussion on building a local AI expert for complex software platforms https://korshunov.ai/es/article/12205-discusion-sobre-la-creacion-de-un-experto-en-ia-local-para-plataformas-de/korshunov.aies2026-07-15T21:17:55.596351+00:00Discusión sobre la creación de un experto en IA local para plataformas de software complejas https://korshunov.ai/ar/article/12204-vulcora-ttlq-thdyan-bqym-51-2k-dwlr-lst-d-lznd-lmkhfy-fy-sb-nmdhj-t-tmd-l/korshunov.aiar2026-07-15T21:08:46.656462+00:00Vulcora تطلق تحدياً بقيمة 51.2K دولار لاستعادة الزناد المخفي في سبعة نماذج تعتمد على SmolLM2 https://korshunov.ai/id/article/12204-vulcora-meluncurkan-tantangan-51-2k-untuk-memulihkan-pemicu-tersembunyi-di-tujuh/korshunov.aiid2026-07-15T21:08:46.656462+00:00Vulcora meluncurkan tantangan $51.2K untuk memulihkan pemicu tersembunyi di tujuh model berbasis SmolLM2 https://korshunov.ai/fr/article/12204-vulcora-lance-un-defi-de-512-k-pour-recuperer-le-declencheur-cache-dans-sept-sur/korshunov.aifr2026-07-15T21:08:46.656462+00:00Vulcora lance un défi de 51,2 K$ pour récupérer le déclencheur caché dans sept modèles basés sur SmolLM2 https://korshunov.ai/ko/article/12204-vulcora-7gae-smollm2-giban-modelyi-sumgyeojin-teurigeo-boggureul-wihan-51-2k/korshunov.aiko2026-07-15T21:08:46.656462+00:00Vulcora, 7개 SmolLM2 기반 모델의 숨겨진 트리거 복구를 위한 $51.2K 챌린지 시작 https://korshunov.ai/ja/article/12204-vulcora-7tsunosmollm2he-sumoterunoyin-saretatorika-hui-fu-notameno51-shi/korshunov.aija2026-07-15T21:08:46.656462+00:00Vulcora、7つのSmolLM2ベースモデルの隠されたトリガー回復のための51.2Kチャレンジを開始 https://korshunov.ai/hi/article/12204-vulcora-ne-saat-smollm2-aadhaarit-monddl-men-chipe-ttrigr-ko-punrpraapt-krne-ke/korshunov.aihi2026-07-15T21:08:46.656462+00:00Vulcora ने सात SmolLM2-आधारित मॉडल में छिपे ट्रिगर को पुनर्प्राप्त करने के लिए $51.2K चुनौती शुरू की https://korshunov.ai/pt-br/article/12204-vulcora-lanca-desafio-de-512k-para-recuperar-gatilho-oculto-em-sete-modelos-no/korshunov.aipt2026-07-15T21:08:46.656462+00:00Vulcora lança desafio de $51,2K para recuperar gatilho oculto em sete modelos baseados no SmolLM2 https://korshunov.ai/zh-hans/article/12204-vulcora-fa-qi-5-12-mo-mei-yuan-tiao-zhan-yi-hui-fu-qi-ge-ji-yu-smollm2-mo-xing/korshunov.aizh-cn2026-07-15T21:08:46.656462+00:00Vulcora 发起 5.12 万美元挑战，以恢复七个基于 SmolLM2 模型中的隐藏触发器 https://korshunov.ai/es/article/12204-vulcora-lanza-un-desafio-de-51-2k-para-recuperar-el-activador-oculto-en-siete-en/korshunov.aies2026-07-15T21:08:46.656462+00:00Vulcora lanza un desafío de $51.2K para recuperar el activador oculto en siete modelos basados en SmolLM2 https://korshunov.ai/ru/article/12204-vulcora-zapuskaet-chellendzh-na-512k-dlia-vosstanovleniia-skrytogo-triggera-v-na/korshunov.airu2026-07-15T21:08:46.656462+00:00Vulcora запускает челлендж на $51,2K для восстановления скрытого триггера в семи моделях на базе SmolLM2