pith. sign in

Hmmt 2025

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it

fields

cs.LG 1

years

2025 1

verdicts

UNVERDICTED 1

representative citing papers

Soft Adaptive Policy Optimization

cs.LG · 2025-11-25 · unverdicted · novelty 4.0

SAPO introduces smooth adaptive gating to replace hard clipping in token- and sequence-level policy optimization for more stable LLM reinforcement learning.

citing papers explorer

Showing 1 of 1 citing paper.

  • Soft Adaptive Policy Optimization cs.LG · 2025-11-25 · unverdicted · none · ref 4

    SAPO introduces smooth adaptive gating to replace hard clipping in token- and sequence-level policy optimization for more stable LLM reinforcement learning.