SFT-then-RL Outperforms Mixed-Policy Methods for LLM Reasoning
Source published ·Modelwire updated
Original coverage: arXiv cs.CL ↗·How Modelwire adds context
The development
A new arXiv paper exposes critical implementation bugs in widely-used LLM training frameworks that have invalidated recent claims about mixed-policy optimization methods. The DeepSpeed optimizer bug silently drops gradient batches during accumulation, while OpenRLHF's loss weighting error compounds the problem, together creating a false performance gap that favors newer techniques over the standard SFT-then-RL baseline. Once corrected, conventional pipelines regain their edge, suggesting the field may have been chasing improvements that don't actually exist. This finding carries immediate implications for practitioners choosing training strategies and raises questions about reproducibility across downstream tools including TRL and Llama-Factory.
Modelwire’s AI-generated summary of coverage from arXiv cs.CL.
Modelwire analysis
ExplainerOur AI-generated reading of the wider context and the next developments to watch.
The deeper issue isn't just that two frameworks had bugs, it's that the bugs were directionally consistent, both systematically disadvantaging the SFT-then-RL baseline, which means any paper benchmarking against that baseline using these tools may have been measuring noise rather than signal.
This connects directly to a reproducibility problem that runs through several recent arXiv findings. The 'Override Gap' paper from the same day (covering hypernetwork-based adaptation) also identified a failure mode that was invisible at the level of reported results, where magnitude mismatches between adapter signals and pretrained priors collapsed accuracy in ways aggregate metrics obscured. Both papers share a common warning: evaluation infrastructure and training pipelines can quietly encode assumptions that corrupt conclusions before a researcher ever looks at a leaderboard. That pattern is worth treating as a category, not a coincidence. The materials discovery and affective computing papers from the same batch don't connect meaningfully here.
Watch whether the authors of the highest-cited mixed-policy papers from 2025 issue corrections or re-run ablations against the patched frameworks within the next 90 days. Silence from that group would itself be informative about how seriously the field treats infrastructure-level reproducibility.
This interpretation is generated from the summary above and the archive coverage cited below. Our methodology · Report an error
Coverage behind this analysis
These archive entries ground the connection in our analysis. They are ordered by source publication date, with links to our coverage and the original sources.
·arXiv cs.LG
The Override Gap: A Magnitude Account of Knowledge Conflict Failure in Hypernetwork-Based Instant LLM Adaptation
Hypernetwork-based adaptation methods like Doc-to-LoRA promise single-pass document internalization into LLMs, but new research exposes a fundamental scaling problem: adapter margins remain constant across inputs while pretrained knowledge margins grow with training frequency, causing accuracy to collapse on high-confidence contradictions. The finding reframes a representational failure as a magnitude mismatch, suggesting that stronger priors systematically…
MentionsDeepSpeed · OpenRLHF · TRL · Llama-Factory · arXiv
How this coverage is produced
Modelwire uses AI to generate summaries and context from source headlines, snippets, and selected archive coverage. Automated checks do not verify every claim, and items are not routinely reviewed by a person before publication. Zacaria Solis operates the site. Read the linked source for the full evidence and report errors through our corrections process.
Modelwire summarizes, we don’t republish. The full content lives on arxiv.org. If you’re a publisher and want a different summarization policy for your work, see our takedown page.