KZ

My project implements Monte Carlo Tree Search - a general inference-time compute technique that has been shown to produce better-quality LLM responses, especially to challenging questions. It can be used with any query. Inspired by the paper https://arxiv.org/pdf/2503.04412