研究团队提出无需验证器的测试时自博弈方法TTSP

1. 13:00
   arXiv 机器学习

   Test-Time Self-Play：无需验证器的测试时扩展方法

   研究人员提出测试时自我博弈方法 Test-Time Self-Play（TTSP），利用非对称自我博弈实现大语言模型在测试阶段的算力扩展，且无需依赖验证器。在难度校准的 LiveCodeBench v6 子集评测中，TTSP 在评估的各类验证器强度下均表现最优；在完全无验证器的极限条件下，TTSP 取得了 mean@8 绝对提升 35% 的成绩。

Current heat 8·Comparable peak 10（Nov 6 14:00）·Comparable change, last 24 hours–

Trends compare only the same parties observed continuously, so the scope can be narrower than current heat counts. Move the pointer or tap the chart to see hourly heat; use the left and right arrow keys on a keyboard.