arXiv cs.CL· Soumye Singhal, Joseph Jennings, Ka Shun Shum, Alexander Bukharin, Khushi Bhardwaj, Zirui Liu, Venkat Srinivasan, Jonathan Li, Yash Jain, Kezhi Kong, Markus Kliegl, Olivier Delalleau, Sanjeev Satheesh, Mostofa Patwary·· 2 天前AI 评分36
为智能体后训练预热的基座模型：一套开放 midtraining 配方
Priming Base Models for Agentic Post-Training: An Open Midtraining Recipe
AI 导读
研究者公开了一套面向智能体后训练的 midtraining 配方，并在 Nemotron 3.5 Lightning 和 Nemotron 3.5 Super 两个模型上验证效果。
正文 · 原文
Authors:Soumye Singhal, Joseph Jennings, Ka Shun Shum, Alexander Bukharin, Khushi Bhardwaj, Zirui Liu, Venkat Srinivasan, Jonathan Li, Yash Jain, Kezhi Kong, Markus Kliegl, Olivier Delalleau, Sanjeev Satheesh, Mostofa Patwary
Abstract:Frontier models, general-purpose and domain-specific alike, now include a midtraining stage between pretraining and post-training, but little has been published about how this stage is built. We present an open midtraining recipe that primes base models for agentic post-training, and we show its effectiveness on two Nemotron models of different sizes. Pretraining typically runs at short context and sees almost no agent interactions, while SFT data is mostly synthetic, and very long SFT stages can hurt the model's later RL trainability. Our midtraining stage sits between the two. It trains at 256K context for hundreds of billions of tokens on a diverse mixture of natural software-engineering data like packed repositories, pull requests, commit histories, and build logs together with a small share of agentic traces. On Nemotron 3.5 Lightning, 500B midtraining tokens improve SWE-bench Verified from 60.7 to 64.5, SWE-bench Multilingual from 55.8 to 62.8, and SWE-bench Pro from 32.0 to 35.5 after identical full SFT. On the larger Nemotron 3.5 Super, 280B midtraining tokens improve DeepSWE, a harder benchmark of long-horizon tasks, from 11.7 to 15.7 and SWE-bench Verified from 72.9 to 74.3 after identical full SFT. Midtraining also improves robustness across agent harnesses on both models. We report the ablations behind the recipe and an evaluation protocol built on short fixed SFT runs, since base-model proxy metrics did not correlate reliably with post-training results.
| Comments: | 19 pages |
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.22019 [cs.CL] |
| (or arXiv:2610.22019v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.22019 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Soumye Singhal [view email]
[v1]
Thu, 15 Oct 2026 14:58:21 UTC (223 KB)
来源：arXiv cs.CL · arxiv.org