GLM-5.2
Z.ai (Zhipu AI)rejection sampling fine-tuning and mask fine-tuning
specific method · filed under post-training
Rejection sampling is a data-filtering method that turns reward functions into filtering criteria to build a high-quality training subset.1 It keeps the highest-quality sample from multiple independently generated candidates, reweighting the training distribution toward better instances.2
Trade-off It improves sample efficiency and training stability, but some filtered-out trajectories still contain valuable high-quality content.3
Summarised by glm-5.3-flash from 1 source; every sentence is checked against the quote it cites.
the reward functions used in RL are transferred into a data filtering pipeline to construct a high-quality training subset— GLM-5: from Vibe Coding to Agentic Engineering
We adopt a Best-of- selection strategy, in which the highest-quality sample is retained from multiple independently generated candidates— GLM-5: from Vibe Coding to Agentic Engineering
leading to improved sample efficiency and enhanced training stability— GLM-5: from Vibe Coding to Agentic Engineering
Also called rejection sampling fine-tuning.
2 quotes from 2 sources, one card per model, each opening with its strongest.
rejection sampling fine-tuning and mask fine-tuning
rejection sampling on those reasoning traces