PreprintarXiv.org2026
SVR: Self-Verifying Refinement via Joint Verdict-Confidence Reinforcement Learning for Adaptive Test-Time Compute
Hongyu Chen, Liang Lin, Guangrun Wang
SVR uses joint verdict-confidence reinforcement learning to train models to self-verify answers, enabling adaptive test-time compute allocation.
0Jul 1, 2026Reinforcement Learning
arXiv