Entropy Ratio Clipping as a Soft Global Constraint for Stable Reinforcement Learning
Paper
•
2512.05591
•
Published
•
16
KlearReasoner
KlearReasoner
-
-
CE-GPPO: Controlling Entropy via Gradient-Preserving Clipping Policy Optimization in Reinforcement Learning
Paper • 2509.20712 • Published • 19 -
Klear-Reasoner: Advancing Reasoning Capability via Gradient-Preserving Clipping Policy Optimization
Paper • 2508.07629 • Published • 42 -
Kwai-Klear/Klear-Reasoner-8B
8B • Updated • 31 • 19 -
Kwai-Klear/KlearReasoner-MathSub-30K
Viewer • Updated • 30k • 115 • 3 -
Kwai-Klear/KlearReasoner-CodeSub-15K
Viewer • Updated • 15k • 153 • 5 -
Kwai-Klear/Klear-Reasoner-8B-SFT
8B • Updated • 12 • 2