1
When Helpfulness Overrides Causal Caution: Context-Dependent Suppression and Recovery in LLMs
当LLM被捧为决策助手时，这份研究揭露了它们“因果谨慎”会在特定语境下悄然失效，是AI安全与可信度的重要警示。
arXiv:2606.24370v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly integrated into decision-support roles in business and p…