作者位次 #2/8ACL 2026Findings· Findings of ACL
Understanding and Mitigating Spurious Signal Amplification in Test-Time Reinforcement Learning for Math Reasoning
Yongcan Yu · Lingxiao He · Jian Liang · Kuangpu Guo · Meng Wang · Qianlong Xie · Xingxing Wang · Ran He
摘要
Test-time reinforcement learning (TTRL) always adapts models at inference time via pseudo-labeling, leaving it vulnerable to spurious optimization signals from label noise.Through an empirical study, we observe that responses with medium consistency form an ambiguity region and constitute the pri...