Efficient Reasoning Exploration via State-Conditioned Latent Steering with Progress Guidance
Best-of-$N$ is a widely used inference strategy for complex reasoning, whose effectiveness depends on whether sampled candidates can cover diverse and high-quality reasoning paths. However, post-trained reasoning models often suffer from \emph{exploration collapse}, where independent rollouts repeatedly follow similar...
Hengyuan Zhang, Chenming Shang, Zunhai Su et al.
· 0 citations