揭示大模型安全训练中“过度拒绝”的成因，用竞争奖励机制平衡安全与可用性。
arXiv:2606.31748v1 Announce Type: new Abstract: Safety training on language models often induces over-refusal: improved safety on harmful prompts at t…
大模型能否自查伦理偏差？新研究引入“良心步骤”用DPO训练实现自我对齐与修正。
arXiv:2606.19527v1 Announce Type: new Abstract: Can Large Language Models (LLMs) discern when their own outputs are misaligned with human ethics? And …
LLM多元性新评估方法：通过潜在视角分析模型多样性的前沿研究。
arXiv:2606.13254v1 Announce Type: new Abstract: The growing need to represent diverse perspectives has increased interest in pluralistic LLM generatio…
LLM也会「助纣为虐」？这篇论文揭露了对话式AI如何被恶意利用放大伤害，并给出缓解方案
arXiv:2606.02423v1 Announce Type: cross Abstract: Large language models (LLMs) can serve as helpful assistants, yet they can equally function as harm …
论文揭示了LLM在回答时隐藏的“盲点偏见”，帮你发现模型没说什么，提升AI对齐与安全认知。
arXiv:2602.10117v5 Announce Type: replace-cross Abstract: Large Language Models (LLMs) often provide chain-of-thought (CoT) reasoning traces that appe…
大模型安全新视角：从提示风险到响应风险的配对分析，揭示LLM安全行为的深层机制。
arXiv:2604.26052v3 Announce Type: replace Abstract: Safety evaluations of large language models (LLMs) typically report binary outcomes, i.e. attack s…
OpenAI详解对齐研究策略：提升AI从人类反馈学习，目标打造能自动解决其余对齐难题的超级对齐系统。
We are improving our AI systems’ ability to learn from human feedback and to assist humans at evaluating AI. Our goal is to build a sufficiently align…