跳到正文
Hugging Face Daily Papers·· 1 天前AI 评分41

Terminal-RSI:通过递归自改写与策略精炼实现终端智能体自我改进

Terminal-RSI: Recursive Self-Improving Terminal Agents via Recursive Self-Rewrite and Policy-Refine

AI 导读

论文提出 Terminal-RSI,交替进行已验证轨迹的递归自改写(RSR)与继承评论家的递归策略精炼(RPR),并引入 TB-GAE、Family Dense Reward 和 Success-Prioritized Reward 等机制。

来源:Hugging Face Daily Papers · huggingface.co