---
id: 20261105-T0-09
title: "RSI-Dojo：训练LLM Agent自动完成LLM后训练"
title_en: "RSI-Dojo Trains LLM Agents to Automate LLM Post-Training"
url: https://ai.daily.yangsir.net/daily/20261105-T0-09
issue_date: 2026-11-05
publish_date: 2026-11-04T05:00:00.000Z
category: research
source_name: "arXiv cs.AI"
source_url: https://arxiv.org/abs/2611.03698
---

# RSI-Dojo：训练LLM Agent自动完成LLM后训练

arXiv论文提出RSI-Dojo，目标是训练Agent自动完成LLM后训练流程。递归自我改进（RSI）指AI系统自主构建和改进模型、harness及底层基础设施，该论文聚焦其中的核心能力：让Agent执行训练任务。目前公开信息仅为摘要，未披露具体基准成绩、模型规模和训练成本。若方法有效，后训练环节的人力投入有望部分由Agent承担。

## English Version

**RSI-Dojo Trains LLM Agents to Automate LLM Post-Training**

An arXiv paper introduces RSI-Dojo, aimed at training agents to automate LLM post-training. Recursive Self-Improvement (RSI) refers to AI systems autonomously building and improving models, harnesses, and underlying infrastructure; the paper targets a central capability—agents that perform training tasks. Only the abstract is available, with no benchmark results, model scale, or training cost disclosed. If the method works, part of the manual effort in post-training could shift to agents.

---

**来源**：[arXiv cs.AI](https://arxiv.org/abs/2611.03698)

**详情页**：https://ai.daily.yangsir.net/daily/20261105-T0-09

---

*智语观潮 · Daily — https://ai.daily.yangsir.net/llms.txt*