AutoTrainess: Teaching Language Models to Improve Language Models Autonomously

arXiv cs.CL Papers

Summary

AutoTrainess presents an LM agent framework that automates the post-training process by externalizing human workflows as explicit interfaces for planning, data preparation, training, evaluation, and logging, achieving significant improvements over CLI-only baselines on PostTrainBench.

arXiv:2606.31551v1 Announce Type: new Abstract: Training language models (LMs) remains a highly human-intensive process, even as frontier language model agents become increasingly capable at software engineering and other long-horizon tasks. A central challenge is that autonomous post-training is not just a coding problem: it requires the agent to repeatedly plan iterations, construct benchmark-aligned data, run stable training jobs, evaluate checkpoints, and preserve experiment state across many hours of interaction. We present AutoTrainess, a LM agent that exposes these operations as a repository of agent-computer interfaces for planning, data preparation, training, evaluation, and logging. Rather than leaving the agent to operate in a raw CLI environment with an underspecified action space, AutoTrainess externalizes prior human experience as explicit workflows, rules, and execution constraints that guide the agent toward effective and reliable training behavior. On PostTrainBench, AutoTrainess consistently outperforms CLI-only baselines, achieving 26.94 average score with GPT-5.4 (Codex) versus 23.21 for CLI-only. It also generalizes across models and harnesses, improving DeepSeek-V4-Flash (OpenCode) from 12.13 to 19.58.
Original Article
View Cached Full Text

Cached at: 07/01/26, 05:34 AM

# AutoTrainess: Teaching Language Models to Improve Language Models Autonomously
Source: [https://arxiv.org/abs/2606.31551](https://arxiv.org/abs/2606.31551)
[View PDF](https://arxiv.org/pdf/2606.31551)

> Abstract:Training language models \(LMs\) remains a highly human\-intensive process, even as frontier language model agents become increasingly capable at software engineering and other long\-horizon tasks\. A central challenge is that autonomous post\-training is not just a coding problem: it requires the agent to repeatedly plan iterations, construct benchmark\-aligned data, run stable training jobs, evaluate checkpoints, and preserve experiment state across many hours of interaction\. We present AutoTrainess, a LM agent that exposes these operations as a repository of agent\-computer interfaces for planning, data preparation, training, evaluation, and logging\. Rather than leaving the agent to operate in a raw CLI environment with an underspecified action space, AutoTrainess externalizes prior human experience as explicit workflows, rules, and execution constraints that guide the agent toward effective and reliable training behavior\. On PostTrainBench, AutoTrainess consistently outperforms CLI\-only baselines, achieving 26\.94 average score with GPT\-5\.4 \(Codex\) versus 23\.21 for CLI\-only\. It also generalizes across models and harnesses, improving DeepSeek\-V4\-Flash \(OpenCode\) from 12\.13 to 19\.58\.

## Submission history

From: Zhaojian Yu \[[view email](https://arxiv.org/show-email/82c66b4c/2606.31551)\] **\[v1\]**Tue, 30 Jun 2026 12:09:51 UTC \(7,262 KB\)

Similar Articles

From Trainee to Trainer: LLM-Designed Training Environment for RL with Multi-Agent Reasoning

arXiv cs.CL

This paper proposes the LLM-as-Environment-Engineer framework, where a policy model analyzes failures to automatically redesign the training environment for reinforcement learning, and introduces MAPF-FrozenLake as a controllable testbed. The framework, using Qwen3-4B, outperforms larger models like GPT and Gemini, showing that policy learning improves the model's ability to diagnose weaknesses.