Google 发布 ReproBridge：提升编程智能体复现 ML 实验成功率
Google 出的 ReproBridge，把复现成功的那次配置打包给下一个 agent 直接用，成功率从 76.4% 拉到 92.6%，跑研究代码的值得一试。
Google 团队推出 ReproBridge，将一次成功的 ML 实验复现打包成可复用资源，包含清洗后的代码、环境配置、实验启动脚本和文字指南。在覆盖 1,029 个任务、来自 36 篇论文的基准测试中，四个模型的复现成功率从 76.4% 提升到 92.6%，API 开销降低 19.7%。最强与最弱模型之间的差距从 21.6 分缩小到 3.9 分。消融测试显示移除代码或环境配置造成的损失最大。每次复现由独立的 auditor 验证运行是否真实发生。
Great paper from Google on making ML experiments reproducible for coding agents. (bookmark it) If you run agents on research code, this one is worth your time. They find that refined artifacts raise agent reproduction success from 76.4% to 92.6% across four models and cut API spending by 19.7%. ReproBridge turns one successful reproduction into a package the next agent can reuse. The package holds the cleaned code, the environment setup, scripts that launch each experiment, and a written guide. A new agent reruns the experiment from that package. A separate auditor then checks that the runs actually happened. The benchmark covers 1,029 tasks from 36 papers, each with five human-verified reference scores. The gap between the best and worst model drops from 21.6 to 3.9 points. Removing the code or the environment setup causes the largest losses. If your agents keep failing to rerun a repo, save the setup steps from the first successful run before trying a bigger model. arxiv.org/abs/2611.08452 Chat with Paper: academy.dair.ai/papers/reprobr… 💬 15 🔄 5 ❤️ 17 👀 1715 📊 17 ⚡
