GIFT matches the optimal policy of GRPO using an endogenous prompt-dependent KL coefficient derived via z-score standardization of implicit rewards.
Musr: Testing the limits of chain-of- thought with multistep soft reasoning, 2024
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
GIFT: Group-Relative Implicit Fine-Tuning Integrates GRPO with DPO and UNA
GIFT matches the optimal policy of GRPO using an endogenous prompt-dependent KL coefficient derived via z-score standardization of implicit rewards.