Explore multiple reasoning paths simultaneously and self-evaluate which branches are most promising — structured deliberation for LLMs.
Five passes over the same idea, each from a different angle. Do them in order, or jump to whichever you need.
Tree-of-Thought (ToT) extends chain-of-thought by generating multiple reasoning branches at each step, evaluating each branch, and pruning unpromising ones. It uses BFS or DFS over a tree of partial solutions. This works exceptionally well for tasks requiring exploration (puzzle-solving, planning, creative writing) where a single reasoning chain might get stuck.
