Tree of thoughts (ToT)
Definition
Tree of Thoughts erweitert CoT durch Pflege mehrerer Argumentationszweige. At each step, the model generates several continuations; a heuristic (or another model) scores them and guides search (z. B. best-first, beam).
Verwenden Sie es, wenn ein einzelnes chain-of-thought path might get stuck (z. B. game moves, multi-step planning) and you can afford multiple LLM calls. It trades compute for better search over the space of solutions. See Schlussfolgern patterns for die vollständige set of options.
Funktionsweise
Start from a root (z. B. die Frage oder der Anfangszustand). Branch: bei jedem Schritt, generate several continuations (z. B. next Schlussfolgern steps or moves). Score each branch with a heuristic or a separate model (z. B. “how promising is this partial solution?”). Expand the best node(s) and repeat; prune low-scoring branches to limit cost. Search strategy (best-first, beam, BFS) and branching factor control exploration vs compute. The tree is built incrementally until a solution is found or a depth/budget limit is reached.
Anwendungsfälle
Tree-of-thoughts is useful wenn Sie want to explore and score multiple solution paths anstatt ein einzelnes chain.
- Game playing and planning where multiple moves need evaluation
- Math or logic with several solution paths to explore
- Creative or Entwurf tasks where generating and scoring options helps
Externe Dokumentation
- Tree of Thoughts (Yao et al.) — ToT paper
- LangChain – Tree of thoughts — ToT and related patterns