研究人员开发了CoTrace框架，用于衡量和揭示人机协作中的目标级贡献，发现虽然AI在整体目标塑造中所占比例较小，但它对具体需求和间接影响有显著贡献。此外，一种名为DGPO的新方法旨在通过解决复杂推理任务中的粗粒度信用分配问题来改进LLM的强化学习。同时，一项关于乌克兰语熵的研究提供了上限并将其与LLM性能进行比较，另一篇论文则探讨了使用稀疏自动编码器进行视觉Transformer的分布外（OOD）检测。
AI
As large language models (LLMs) increasingly shape how users form, refine, and extend their goals, attributing contributions in human-AI collaboration becomes critical for users calibrating their own reliance and for evaluators assessing AI-assisted work. Yet existing methods foc…
As large language models (LLMs) increasingly shape how users form, refine, and extend their goals, attributing contributions in human-AI collaboration becomes critical for users calibrating their own reliance and for evaluators assessing AI-assisted work. Yet existing methods foc…
arXiv cs.AI
TIER_1English(EN)·Pacome Simon Mbonimpa·
arXiv:2605.02285v1 Announce Type: new Abstract: The deployment of Large Language Models (LLMs) for specialized engineering domains, such as circuit analysis, often faces a trade-off between reasoning accuracy and computational efficiency. Traditional evaluation methods treat mode…
arXiv:2605.03327v1 Announce Type: new Abstract: Reinforcement learning is crucial for aligning large language models to perform complex reasoning tasks. However, current algorithms such as Group Relative Policy Optimization suffer from coarse grained, sequence level credit assign…
Reinforcement learning is crucial for aligning large language models to perform complex reasoning tasks. However, current algorithms such as Group Relative Policy Optimization suffer from coarse grained, sequence level credit assignment, which severely struggles to isolate pivota…
arXiv:2604.27534v1 Announce Type: new Abstract: In natural language processing, the entropy of a language is a measure of its unpredictability and complexity. The first study on this subject was conducted by Claude Shannon in 1951. By having participants predict the next characte…
In natural language processing, the entropy of a language is a measure of its unpredictability and complexity. The first study on this subject was conducted by Claude Shannon in 1951. By having participants predict the next character in a sentence, he was able to approximate the …
Sparse Autoencoders (SAEs) have demonstrated significant success in interpreting Large Language Models (LLMs) by decomposing dense representations into sparse, semantic components. However, their potential for analyzing Vision Transformers (ViTs) remains largely under-explored. I…
arXiv:2601.09093v2 Announce Type: replace Abstract: Large Language Models (LLMs) can enhance reasoning capabilities through test-time scaling by generating multiple traces. However, the combination of lengthy reasoning traces with multiple sampling introduces substantial computat…
arXiv cs.LG
TIER_1English(EN)·Achref Jaziri, Martin Rogmann, Martin Mundt, Visvanathan Ramesh·
arXiv:2510.17381v2 Announce Type: replace Abstract: Detecting out-of-distribution (OOD) data is critical for machine learning, be it for safety reasons or to enable open-ended learning. However, beyond mere detection, choosing an appropriate course of action typically hinges on t…
arXiv:2604.26409v1 Announce Type: new Abstract: Sparse Autoencoders (SAEs) have demonstrated significant success in interpreting Large Language Models (LLMs) by decomposing dense representations into sparse, semantic components. However, their potential for analyzing Vision Trans…
Sparse Autoencoders (SAEs) have demonstrated significant success in interpreting Large Language Models (LLMs) by decomposing dense representations into sparse, semantic components. However, their potential for analyzing Vision Transformers (ViTs) remains largely under-explored. I…
