← back to paper
arxiv: 2605.17694 · 2 revisions
Do LLM Agents Mirror Socio-Cognitive Effects in Power-Asymmetric Conversations?