This study investigates how semantically similar concerns presented through different contextual framings elicit varying responses from instruction-tuned large language models, potentially challenging system reliability. Using controlled matched prompts and layer-wise probing analyses, the authors demonstrate that framing systematically alters interpretive response tendencies across multiple model architectures.
Researchers propose ReaORE, a framework for open relation extraction that utilizes large reasoning models to achieve reliable generalization to unseen relation types. The method addresses limitations of current clustering and direct generation approaches through a coarse-to-fine reasoning process.
This study investigates the presence and structure of emotion vectors in open-weight large language models, specifically Apertus-8B-Instruct-2509 and Gemma-4-E4B-it. The research confirms that these models encode valence geometry with high correlation to human psychological structures, approaching the levels previously observed in Claude Sonnet 4.5.
The authors introduce MinGram, a minimalist unigram tokenizer that simplifies training by using a BPE-derived seed vocabulary, Hard EM on a minimum-token path, and a single flat score-pruning step. This approach removes the need for suffix arrays, forward-backward passes, and iterative prune loops, making the procedure significantly less complex than standard methods.
This work addresses the tendency of multimodal large language models to produce overconfident outputs in Medical Visual Question Answering by proposing a training-based framework that finetunes these models for better calibration. The method employs a composite loss function combining Brier-style calibration, anchor regularization, contrastive image-text alignment, and KL divergence terms to align model confidence with actual correctness.
Researchers propose Psy-CoT, a psychology-grounded chain-of-thought framework that decomposes pre-response reasoning into Interaction Perception, Psychological Empathy, and Logical Construction to improve character fidelity. To address gradient misalignment in reinforcement learning, they introduce Role-Aware Policy Optimization (RAPO), which uses profile-token mutual information to weight gradients asymmetrically.
Researchers introduce NuclearQAv2, a new benchmark designed to assess the reliability of large language models in nuclear engineering by testing factual knowledge, quantitative reasoning, and conceptual understanding.
Researchers propose a Judge-Aware Gated Multi-Task Learning architecture that disentangles objective case facts from adjudicative context to improve legal outcome prediction. The model uses a fine-grained outcome taxonomy and a gated fusion mechanism to dynamically modulate reliance on judge identity, evaluated on 13,937 UK Employment Tribunal decisions.
A study introduces the "riddle riddle" paradigm to determine whether large language models (LLMs) rely on flexible reasoning or pattern matching, revealing that humans and LLMs fail in opposite directions. In experiments involving nine state-of-the-art LLMs and 100 human participants, LLMs performed significantly worse on riddle riddles than on genuine riddles, while humans showed the reverse trend.
Researchers introduce HarmVideoBench, a multi-layered diagnostic benchmark designed to evaluate large vision-language models on their ability to understand harmful videos beyond superficial cues. The benchmark addresses limitations in existing works by incorporating explanatory rationales and assessing three hierarchical dimensions of harm: Observable Evidence, Clip-Internal Meaning, and Beyond-Clip Reasoning.
This study applies Large Language Models to forecasting tasks and uses sparse autoencoders to analyze their internal states, distinguishing between time-specific knowledge and generalizable patterns. The research identifies specific features associated with both time-aware reasoning and look-ahead-biased reasoning.
The article proposes Syntactic Belief Update, a model that predicts processing difficulty in garden path sentences by measuring the magnitude of syntactic belief updates via generalized Rényi divergence. This approach outperforms lexical surprisal by providing a better fit to human reading time data.
The authors introduce AIMS, a dataset of 1,724 human-annotated difficult safety prompts paired with intent descriptions and harm labels, to evaluate intent-aware training across multiple regimes. They argue that modeling user intent as an explicit signal significantly improves the robustness of safety classifiers.
The authors propose BINEVAL, a framework that decomposes evaluation criteria into atomic binary questions to provide interpretable, multi-dimensional scores for large language models. This approach generates transparent question-level feedback and calibrated overall scores by having an LLM answer fine-grained evaluation questions independently for each output.
The datasette-export-database plugin version 0.3a2 addresses a compatibility issue caused by an overly strict dependency constraint in the previous release.
The llama.cpp b9827 release introduces a performance optimization for CUDA by adding a cudaMemcpy2DAsync fast path to the ggml_cuda_cpy function. This change accelerates same-type, same-shape strided copies where tensors are not fully contiguous but each row is contiguous, replacing slower element-wise scalar copy kernels.
The author introduces BatonBot, an open-source local-first application designed to streamline AI coding workflows by reducing the need for constant user supervision. The tool addresses the inefficiency of sequential agent interactions by allowing users to set up tasks and track progress visually on a Kanban-style board.
The open-source project audio.cpp provides a native C++ inference framework for audio models built on top of ggml, currently supporting 12 released model families including TTS, ASR, and voice conversion. Benchmarks on Ubuntu/CUDA demonstrate that text-to-speech performance in this runtime is up to 5x faster than the corresponding Python reference implementations.
Bruce Schneier discusses a recent German ruling that holds Google liable for errors in its AI overviews, arguing that AI agents should be treated as agents of the deploying organization.
JetSpec introduces a speculative decoding method called causal parallel tree drafting that co-optimizes drafting cost and quality to reduce LLM generation latency. The approach achieves up to 9.64x end-to-end speedup on MATH-500 and 4.58x on open-ended chat while maintaining lossless accuracy.