SAPO introduces smooth adaptive gating to replace hard clipping in token- and sequence-level policy optimization for more stable LLM reinforcement learning.
Measuring multimodal mathematical reasoning with math-vision dataset
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
-
Soft Adaptive Policy Optimization
SAPO introduces smooth adaptive gating to replace hard clipping in token- and sequence-level policy optimization for more stable LLM reinforcement learning.