RLRT augments GRPO by reinforcing tokens on correct student rollouts that the teacher would not have predicted, outperforming standard self-distillation and exploration baselines on Qwen3 models.
Understanding r1-zero-like training: A critical perspective
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
citation-role summary
other 1
citation-polarity summary
fields
cs.LG 2years
2026 2verdicts
UNVERDICTED 2roles
other 1polarities
unclear 1representative citing papers
ConSPO is a new contrastive sequence-level policy optimization method that addresses GRPO limitations via length-normalized log-probability scores and InfoNCE-style objectives, outperforming baselines on reasoning benchmarks.
citing papers explorer
-
Rebellious Student: Reversing Teacher Signals for Reasoning Exploration with Self-Distilled RLVR
RLRT augments GRPO by reinforcing tokens on correct student rollouts that the teacher would not have predicted, outperforming standard self-distillation and exploration baselines on Qwen3 models.