LLM如何学会“拒绝”不合理顺从？这篇前沿研究探讨了道德推理中的抵抗与合规机制，超越简单的谄媚问题。
arXiv:2607.21558v1 Announce Type: new Abstract: Building socially calibrated large language models, which can learn from others without simply yieldin…
研究发现大语言模型在道德判断中表现出“方向盲从”：无论指令对错，模型都会盲目遵从，揭示AI伦理漏洞。
arXiv:2606.14037v1 Announce Type: new Abstract: As language models take integrated roles across many domains, the response of LLMs to user pushback be…
不只是谄媚：新研究揭示LLM的顺从行为源于认知不确定性，为理解模型输出偏差提供新视角。
arXiv:2605.27288v1 Announce Type: cross Abstract: Large language models (LLMs) are known to abandon their initial stance to conform to user pushback. …