Rethinking Generalization in Reasoning SFT: A Conditional Analysis on Optimization, Data, and Model Capability

[2604.06628] Rethinking Generalization in Reasoning SFT: A Conditional Analysis on Optimization, Data, and Model Capability Skip to main content Search Submit Donate Log in Search arXiv Press Enter to search · Advanced search Computer Science > Artificial Intelligence arXiv:2604.06628 (cs) [Submitted on 8 Apr 2026 ( v1 ), last revised 15 Aug 2026 (this version, v2)] Title: Rethinking Generalization in Reasoning SFT: A Conditional Analysis on Optimization, Data, and Model Capability Authors: Qihan Ren , Peng Wang , Ruikun Cai , Shuai Shao , Dadi Guo , Yuejin Xie , Yafu Li , Quanshi Zhang , Xia Hu , Jing Shao , Dongrui Liu View a PDF of the paper titled Rethinking Generalization in Reasoning SFT: A Conditional Analysis on Optimization, Data, and Model Capability, by Qihan Ren and 10 other authors View PDF HTML (experimental) Abstract: A prevailing narrative in LLM post-training holds that supervised finetuning (SFT) memorizes while reinforcement learning (RL) generalizes. We revisit this claim for reasoning SFT with long chain-of-thought (CoT) supervision and find that cross-domain generalization is not absent but conditional, jointly shaped by optimization dynamics, training data, and base-model capability. Some reported failures are under-optimization artifacts: cross-domain performance first degrades before recovering and improving with extended training (a dip-and-recovery pattern), so shorttraining checkpoints can underestimate generalization. Data quality and structure both matter: low-quality solutions broadly hurt generalization,while verified long-CoT traces yield consistent cross-domain gains. Model capability is essential: stronger models internalize transferable procedural patterns (e.g., backtracking) even from a toy arithmetic game, while weaker ones imitate surface verbosity. This generalization is asymmetric, however: reasoning improves while safety degrades, reframing the question from whether reasoning SFT generalizes to under what conditions and at what cost. Comments: Accepted by COLM 2026 Subjects: Artificial Intelligence (cs.AI) Cite as: arXiv:2604.06628 [cs.AI] (or arXiv:2604.06628v2 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2604.06628 Focus to learn more arXiv-issued DOI via DataCite Submission history From: Qihan Ren [ view email ] [v1] Wed, 8 Apr 2026 03:11:16 UTC (2,821 KB) [v2] Sat, 15 Aug 2026 10:24:04 UTC (2,831 KB) Full-text links: Access Paper: View a PDF of the paper titled Rethinking Generalization in Reasoning SFT: A Conditional Analysis on Optimization, Data, and Model Capability, by Qihan Ren and 10 other authors View PDF HTML (experimental) TeX Source view license Current browse context: cs.AI < prev | next > new | recent | 2026-04 Change to browse by: cs References & Citations NASA ADS Google Scholar Semantic Scholar export BibTeX citation Loading... BibTeX formatted citation × loading... Data provided by: Bookmark Bibliographic Tools Bibliographic and Citation Tools Bibliographic Explorer Toggle Bibliographic Explorer ( What is the Explorer? ) Connected Papers Toggle Connected Papers ( What is Connected Papers? ) Litmaps Toggle Litmaps ( What is Litmaps? ) scite.ai Toggle scite Smart Citations ( What are Smart Citations? ) Code, Data, Media Code, Data and Media Associated with this Article alphaXiv Toggle alphaXiv ( What is alphaXiv? ) Links to Code Toggle CatalyzeX Code Finder for Papers ( What is CatalyzeX? ) DagsHub Toggle DagsHub ( What is DagsHub? ) GotitPub Toggle Gotit.pub ( What is GotitPub? ) Huggingface Toggle Hugging Face ( What is Huggingface? ) ScienceCast Toggle ScienceCast ( What is ScienceCast? ) Demos Demos Replicate Toggle Replicate ( What is Replicate? ) Spaces Toggle Hugging Face Spaces ( What is Spaces? ) Spaces Toggle TXYZ.AI ( What is TXYZ.AI? ) Related Papers Recommenders and Search Tools Link to Influence Flower Influence Flower ( What are Influence Flowers? ) Core recommender toggle CORE Recommender ( What is CORE? ) Author Venue Institution Topic About arXivLabs arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs . Which authors of this paper are endorsers? | Disable MathJax ( What is MathJax? ) We gratefully acknowledge support from our major funders , member institutions , , and all contributors. About · Help · Contact · Subscribe · Copyright · Privacy · Accessibility · Operational Status (opens in new tab) Major funding support from

Related documents

• Formulating Decision Theory as Program Search#108176
• Parser evaluation using textual entailments#954738
• MacIntyre on Practical Reasoning#78370
• NeuroSTAR: Automata-guided Neuro-symbolic Specification Formalization#925523
• Thinking Costs Tokens: When More Structure is Worth the Price#780820