AutoTrainess：教会语言模型自主改进语言模型

摘要

训练语言模型（LM）仍然是一个高度依赖人力的过程，即使前沿语言模型代理在软件工程及其他长周期任务中能力日益增强。核心挑战在于，自主后训练不仅仅是编码问题：它要求代理反复规划迭代、构建与基准对齐的数据、运行稳定的训练任务、评估检查点，并在数小时的交互过程中保存实验状态。我们提出 AutoTrainess，这是一个语言模型代理，它将上述操作暴露为代理-计算机接口的仓库，用于规划、数据准备、训练、评估和日志记录。AutoTrainess 并非让代理在原始命令行界面环境（CLI）中操作行动空间不明确的任务，而是将先前的人类经验外化为明确的工作流、规则和执行约束，引导代理实现有效且可靠的训练行为。在 PostTrainBench 上，AutoTrainess 持续优于仅使用 CLI 的基线，在 GPT-5.4 (Codex) 上获得 26.94 的平均分数，而 CLI 基线仅为 23.21。它还跨模型和工具集实现泛化，将 DeepSeek-V4-Flash (OpenCode) 的分数从 12.13 提升至 19.58。

English

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.