1 paper · 1 filter
Yanxi Chen, Xuchen Pan, Yaliang Li +2
We propose two simple, principled and practical algorithms that enjoy provable scaling laws for the test-time compute of large language models (LLMs). The first one is a two-stage…