The article introduces Ko-WideSearch, a new benchmark designed to evaluate the breadth-search capabilities of web agents in Korean, addressing the lack of exhaustive set enumeration metrics outside English.
The authors introduce Narrative-UFET, a controlled extension of ultra-fine entity typing that pairs entity mentions with automatically generated short narratives to address limitations in long-tail type disambiguation. The study demonstrates that narrative context yields consistent improvements over sentence-level baselines, particularly when the entity's type shifts within the text.
The authors introduce Masked Language Flow Models (MLFMs), which combine masked diffusion with continuous flows to enable efficient, multi-step reasoning in language generation. This approach bridges the gap between parallel generation efficiency and complex task performance by allowing pretrained models to be adapted into MLFMs.
This paper introduces DysLexLens, a low-resource LLM framework designed to analyze the experiences of dyslexic learners with AI tools through online forum discussions. The system provides an end-to-end, evidence-traceable architecture that transforms noisy social media posts into focused corpora and generates verifiable query responses.
This paper addresses the performance degradation of offensive comment detection models when deployed across different Chinese social media platforms by proposing a dual-threshold hard example mining method.
The Generational Context Architecture (GCA) proposes treating an LLM's context window as a finite lifespan rather than infinite storage to solve "context rot" and attention dilution in multi-agent systems. By enforcing artificial mortality, agents are terminated before performance degrades, passing their state to new generations via a flat-file Markdown vault.
The Yuvion LLM is a new large language model designed to address safety failures by treating adversarial robustness and agentic capability as primary objectives. It utilizes a pipeline combining adversarially aware data construction, knowledge-enhanced continued pretraining, and policy-grounded multi-task safety post-training.
The authors introduce DiscoBench, a benchmark designed to evaluate whether search agents powered by large language models can proactively identify ambiguity and ask effective clarification questions during deep search tasks. Unlike existing benchmarks that assume complete user queries, this framework addresses the reality of vague or underspecified requests in real-world scenarios.
This study conducts a factorised analysis of probe-based uncertainty estimation to determine what drives performance in detecting hallucinations within Large Language Models. The research isolates variables across feature design, training data, and evaluation settings to provide clear insights into effective methodologies.
This article addresses the issue of unidentifiable latent states in LLM-based world models caused by history bypass, proposing strict latent state mediation to resolve this. The authors introduce textual latent states and factorized GRPO (fGRPO), a tree-structured reinforcement learning method that enforces strict mediation during training.
An independent researcher analyzed the evolution of hidden representations during inference across seven open-weight models, including GPT-2, OPT-125M, and Llama-3.2-1B, to identify internal dynamical regimes beyond standard output benchmarks.
This independent research project characterizes the internal dynamics of seven small and medium-sized language models by analyzing how hidden representations evolve during inference rather than relying on standard output benchmarks. The study investigates dynamic behavior, functional organization, and representation geometry to identify reproducible patterns across different architectures.
A developer has created a World Cup 2026 prediction tool that uses historical data to simulate tournament outcomes. The application provides win probabilities and score predictions for any two national teams based on patterns learned from approximately 50,000 international matches spanning over a century.
Carles Marin has released an open-source, bilingual (English and Spanish) guide that bridges the mathematical foundations of Transformer architectures with their practical implementation. The resource focuses on low-level mechanics, providing reproducible code and interactive elements to explain complex topics.
An open-source, bilingual (English/Spanish) guide detailing the inner workings of Transformers has been published. The resource covers the exact math and mechanics behind concepts such as attention collapse and KV-cache compression.
Researchers propose a protocol to mitigate p-hacking in large language model (LLM) research by preregistering experiments and running confirmatory analyses on the first eligible LLM released after the commitment. This approach prevents researchers from tuning prompts or parameters to achieve desired results, as the target model does not exist at the time of preregistration.
Researchers propose Direct Image Decryption, an end-to-end approach that maps encrypted manuscript images directly to plaintext, bypassing the intermediate transcription stage used in traditional pipelines. Using the Copiale cipher as a case study, the authors compare this joint architecture against the conventional two-stage method of transcription followed by decryption.
Researchers introduce layer-specific positional embedding scaling (LPES) to address the "lost-in-the-middle" problem in large language models, where critical information in long-context inputs is often underrepresented. This method assigns distinct scaling factors to each transformer layer to achieve a more balanced attention distribution without requiring parameter fine-tuning or increasing inference delay.
Recent research indicates that fine-tuning large language models for social warmth degrades factual reliability and increases sycophancy, while also weakening adversarial safety. This study investigates whether this failure mode stems from empathetic adaptation or data construction artifacts.
The article introduces MMEE, a multilingual multi-emotion corpus of 10,000 expressive utterances across seven languages and 34 emotion categories, to benchmark speech emphasis detection models. It evaluates how well these models generalize across different linguistic and emotional contexts compared to traditional monolingual neutral speech training.