Sakana AI j made it possible for LLMs to reason with both depth AND breadth !!

This post introduces Adaptive Monte Carlo Tree Search (AB-MCTS), a novel method that allows Large Language Models (LLMs) to reason with both depth and breadth. It explains the problem with current LLM strategies, details how AB-MCTS works, presents its impressive results across various benchmarks, and highlights its potential for future AI collaboration.