{"version":1,"type":"story","url":"https://digestai.news/story/researchers-introduce-rsi-dojo-to-train-llm-agents-for-post-training","json":"https://digestai.news/story/researchers-introduce-rsi-dojo-to-train-llm-agents-for-post-training.json","markdown":"https://digestai.news/story/researchers-introduce-rsi-dojo-to-train-llm-agents-for-post-training.md","slug":"researchers-introduce-rsi-dojo-to-train-llm-agents-for-post-training","headline":"Researchers introduce RSI-Dojo to train LLM agents for post-training","summary":"A new paper from arXiv introduces RSI-Dojo, a collection of 281 executable model training environments designed to teach large language model agents how to conduct post-training autonomously. The work addresses a gap in Recursive Self-Improvement research, where previous studies focused on workflow optimization rather than embedding the ability to train models directly into the agent itself.\n\nTo build this environment, the authors developed RSI-Craft, an automatic pipeline that converts human research artifacts into executable tasks. This pipeline collects papers from top venues, standardizes datasets and metrics, and filters tasks based on data size and reliability. The team then used frontier models to collect high-quality training trajectories, which were refined through rubric-based and outcome-based filtering.\n\nThe researchers fine-tuned Qwen3.5-9B and Qwen3.6-35B-A3B on these trajectories. The resulting models, RSI-Dojo-9B and RSI-Dojo-35B-A3B, showed improvements of 8.74 and 16.87 points on PostTrainBench, respectively. Notably, the 35B variant reached a score of 29.95 on PostTrainBench and 53.29 on PostTrainBench-Lite, matching the performance of frontier models like GLM-5.2 and GPT 5.5 in this specific task.","keyPoints":["RSI-Dojo provides 281 executable environments for teaching LLM agents to perform post-training tasks.","Fine-tuned Qwen models improved by 8.74 and 16.87 points on PostTrainBench using RSI-Dojo trajectories.","RSI-Dojo-35B-A3B matched frontier models like GPT 5.5 on PostTrainBench-Lite with a score of 53.29."],"whyItMatters":"This research advances Recursive Self-Improvement by enabling agents to handle complex model training tasks, potentially reducing the human effort required for iterative model development and improving the efficiency of AI research pipelines.","category":{"slug":"research","name":"Research","url":"https://digestai.news/category/research"},"entities":{"companies":[],"models":["Qwen3.5-9B","Qwen3.6-35B-A3B","GLM-5.2","GPT 5.5"],"people":[]},"firstPublishedAt":"2026-11-04T05:00:00Z","updatedAt":"2026-11-04T05:00:00Z","sourceCount":1,"hasPrimarySource":true,"sources":[{"outlet":"arXiv cs.AI","title":"Training LLM Agents to Automate LLM Post-Training with RSI-Dojo","url":"https://arxiv.org/abs/2611.03698","publishedAt":"2026-11-04T05:00:00Z","type":"primary","primary":true,"lead":true}],"sourceNotes":null,"discussions":[],"thread":{"title":"AI Agents Rise Across Finance and Science","url":"https://digestai.news/thread/navalia-improves-cowork-agent-performance-over-qwen3-6-35b-a3b-on-finance","storyCount":3},"cite":{"text":"Digest AI, \"Researchers introduce RSI-Dojo to train LLM agents for post-training\", 4 November 2026, https://digestai.news/story/researchers-introduce-rsi-dojo-to-train-llm-agents-for-post-training","publisher":"Digest AI","title":"Researchers introduce RSI-Dojo to train LLM agents for post-training","datePublished":"2026-11-04T05:00:00Z","url":"https://digestai.news/story/researchers-introduce-rsi-dojo-to-train-llm-agents-for-post-training"},"generatedBy":"Written by Digest AI's editorial model from the linked sources; the sources are the record.","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse"}