More Correct Mass, Worse Answers: Why Power Sampling Can Fail and How to Fix It

Excerpt

arXiv:2608.14420v1 Announce Type: new Abstract: Power Sampling sharpens a language model's distribution over complete generation trajectories, offering a verifier-free way to improve reasoning at inference time. It also has the potential to serve as a general-purpose front end for a broad range of downstream sampling methods. However, we uncover a striking paradox: Power Sampling can drive more probability mass toward correct trajectories while degrading the downstream inference it is intended t