1
Mind the Sim2Real Gap in User Simulation for Agentic Tasks
用户模拟能否真实反映人类行为？这篇论文直击Agent评估中的Sim2Real痛点，做多轮交互评估必读。
arXiv:2603.11245v2 Announce Type: replace Abstract: As NLP evaluation shifts from static benchmarks to multi-turn interactive settings, LLM-based simu…