Revising Context, Shifting Simulated Stance: Auditing LLM-Based Stance Simulation in Online Discussions 事件

Revising Context, Shifting Simulated Stance: Auditing LLM-Based Stance Simulation in Online Discussions arXiv:2606.06443v1 Announce Type: new Abstract: Large language models are increasingly used to simulate social media users and infer how individuals may respond to online discussions. However, it remains unclear whether these simulations reflect precise user-specific beliefs or whether they are highly sensitive to semantically independent changes in conversational contexts. In this work, we s