Recursa 论文提出递归自我改进框架，在 Firm Knowledge 基准上提升通过率
Recursa: Toward Recursively Self-Improving Foundation Model Systems
一篇关于让模型在没老师监督下自己变强的方法论文，用 DeepSeek-V4.1-Flash 测的，全通过率从 30% 拉到 55%，做 agent 自我改进的可以看看。
arXiv 论文介绍 Recursa，一个不依赖人类专家或更强教师模型的自我改进基础模型系统。系统通过知识仓库存储领域技能、应用时机说明以及过往任务记忆和反思。工作流程是先从领域材料自举生成仓库，再由并行学习者执行任务、综合局部改进，验证后合并进仓库。在 Harvey LAB 的 Firm Knowledge 基准上，Recursa 将 DeepSeek-V4.1-Flash 的 held-out 全通过率从 30.0% 提到 55.0%，Qwen3.8-Flash 从 35.0% 提到 46.0%。
Recursa: Toward Recursively Self-Improving Foundation Model Systems
Foundation models in changing environments can encounter unfamiliar domain knowledge that their original training did not cover. However, new domains rarely provide large amounts of expert-created training data or reliable supervisory signals for further training. We introduce Recursa, a foundation model system that adapts to new domains by learning from its own experience without human experts or a stronger teacher model. Recursa uses a knowledge repository to store what it learns: domain-specific skills and instructions on when to apply them, along with memories of past task experiences and reflections on what to improve. Recursa bootstraps the repository from domain materials and then refines it through recursive self-improvement. It uses parallel learners to execute tasks and synthesize local improvements, then validates and merges the improvements into the knowledge repository. On Harvey LAB's Firm Knowledge benchmark, Recursa raises the all-pass rate on held-out test tasks from 30.0% to 55.0% with DeepSeek-V4.1-Flash and from 35.0% to 46.0% with Qwen3.8-Flash.