Replacing the importance sampling ratio with a stop-gradient self-anchored ratio for positive advantages yields unclipped, REINFORCE-equivalent gradients that improve exploration without training instability.
American invitational mathematics examination-aime 2024
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.LG 2years
2026 2representative citing papers
EDAS modulates RL advantage signals for incorrect rollouts by amplifying penalties on repeated errors and attenuating them on rare ones, yielding average gains of 6.29 points over DAPO on Qwen3-8B across seven math benchmarks.
citing papers explorer
-
UP: Unbounded Positive Asymmetric Optimization for Breaking the Exploration-Stability Dilemma
Replacing the importance sampling ratio with a stop-gradient self-anchored ratio for positive advantages yields unclipped, REINFORCE-equivalent gradients that improve exploration without training instability.
-
Leveraging Error Diversity in Group Rollouts for Reinforcement Learning
EDAS modulates RL advantage signals for incorrect rollouts by amplifying penalties on repeated errors and attenuating them on rare ones, yielding average gains of 6.29 points over DAPO on Qwen3-8B across seven math benchmarks.