AI 犯错后，终于有人把错误本身当教材
大模型 Agent 失败后，通常只留下一个「没成功」的分数，过程全被扔掉。这篇论文把 5 万条真实失败过程——模型当时看到了什么、选了哪步、环境怎么回应——连同人工诊断和修正建议一起存成数据集，再拿它训练模型。结果很直接：让模型先学「自己错在哪」再学「怎么改」，在 3 千多组对照测试里，修正方案通过率从 18.4% 提到 51.1%;用完整诊断微调，模型判断「哪一步该改」的准确率从 47.2% 涨到 63.6%，超过最强的提示词方案。这不是你明天能用的产品，但它指向一个更本质的思路：AI 的进步不只在「做对更多」，也在「知道自己怎么错的」。
📄 原文摘要(英文)
An unsuccessful LLM agent rollout contains more information than its final reward: the observations available to the agent, the actions it chose, and the environment's responses. Reusing this experience for learning requires identifying a decision to revise and testing a concrete alternative. We introduce the Agent Error Dataset (AED), comprising 50,228 error-diagnosis pairs from 9,961 source tasks across 33 environments, 19 harness families, and 23 policy models in text-based agent systems. We retain source traces and execution metadata to support cross-setting failure analysis and re-diagnosis without repeating the original rollout. Our five-stage Agentic Error-to-Training (AET) pipeline collects natural failures, generates diagnoses and proposed corrections, and checks them against recorded evidence. Where replay is supported, we compare corrections with original-action retries from the same checkpoint under matched execution settings. We then construct separate training views for diagnosis and actor recovery. Across 3,062 matched replay pairs, first-proposal corrections raise verifier pass rates from 18.4% to 51.1%, a gain of 32.7 percentage points. Using a separately frozen diagnosis release, full-diagnosis fine-tuning on 1,656 source tasks raises Qwen3-8B's exact-step agreement with internal teacher labels from 47.2% to 63.6%, averaged over three seeds on a 943-case holdout. The strongest prompted reference in this comparison scores 54.7%, and mean agreement improves at each of four increasing training-set sizes. In a single-seed comparison of actor-training recipes, action-only repair training scores 6.67 percentage points higher on WebShop-lite than success-only training.