新基准CORTEX专为3D胸部CT的多模态大语言模型设计，强调结构化推理而非仅看最终答案，让诊断结果可溯源至证据，提升可信度。
arXiv:2606.27264v1 Announce Type: new Abstract: Reasoning in multimodal large language models (MLLMs) has shown strong promise in medical imaging. How…
一次改写质量审计揭露了数学推理基准中语义漏洞，排名翻盘只需移除3.1%错误项，单一模型评估的盲区由此暴露。
arXiv:2605.29001v1 Announce Type: cross Abstract: A paraphrase-quality audit of MathCheck (ICLR 2025) detected 4 semantically incorrect paraphrases in…
揭示长上下文LLM因位置偏差导致推理失败的盲点，挑战现有基准评估体系。
arXiv:2605.23170v1 Announce Type: cross Abstract: Position-controlled evaluation is standard for retrieval tasks such as Needle-in-a-Haystack and RULE…
新基准MixRea评估LLM在显式与隐式推理上的表现，揭示推理能力的短板。
arXiv:2605.20128v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly integrated into high-stakes decision-making. Inspired by…