Comparing and selecting task-oriented LLM agents increasingly relies on a low-cost offline evaluation gate: persona-driven LLM user-simulators converse with …
机构:Amazon
来源:arXiv 2609.12191 | AI4Papers 论文推荐平台