TRACE Bench：任务驱动的角色扮演智能体检查表评估框架

AI 摘要

TRACE Bench 是一个任务驱动的角色扮演智能体检查表评估框架，将角色档案分解为固定检查表，并通过用户智能体自然对话动态更新检查表状态，使评分可追溯到具体检查项和对话证据。实验显示，其覆盖率在更少轮次内达到 99.91%，而现有 M2 自由对话转录仅覆盖 73.74%。该框架支持 26 个模型的排名、能力分解和闭环基准演化，以提升评估的可审计性和可靠性。

正文节选

GameMind TRACE Bench: Task-driven Roleplay Agentic Checklist Evaluation Roleplay evaluation should do more than assign a single score: it should reveal which role requirements were tested, which failed, and which dialogue evidence supports the judgment. We propose TRACE Bench, a task-driven agentic checklist evaluation framework. It decomposes each role profile offline into a fixed checklist, then uses a User Agent to converse naturally with the target roleplay model while privately updating che

发布时间：2026-08-13 12:00

抓取时间：2026-08-13 12:05

来源机构：arXiv