trends.barnas.net

method · first seen 5 Oct 2026

rejection-sampling fine-tuning

Also: RFT, rejection-sampled trajectories

Training a model on its own sampled outputs that pass a correctness check, keeping only successful samples.

Mentions per week

0219 May11 Aug10 NovPapersCodeDiscussion

PapersCodeDiscussion

Earliest sources

Latest sources

Mentioned together with

group relative policy optimizationpass@kcoding benchmark generationrecursive self-improvementlow-rank adaptationreinforcement learning with verifiable rewardsoverthinking in reasoning models