Tree of Thoughts: Deliberate Problem Solving with Large Language Models
2023 — branching search over intermediate thoughts (NeurIPS)
Tree-of-thought prompting — explore, evaluate, and backtrack over coherent text thoughts as search nodes.
Yao, S., Zhao, J., Yu, D., Du, N., Shafran, I., Narasimhan, K., & Cao, Y. (2023). Tree of Thoughts: Deliberate Problem Solving with Large Language Models. Advances in Neural Information Processing Systems 36 (NeurIPS 2023); arXiv:2305.10601.
Why we cite it
Defines Tree of Thoughts (ToT) — tree topology over legible intermediate steps, with explicit reasoning schedule (how the tree is traversed). See taxonomy in Demystifying Chains, Trees, and Graphs of Thoughts and Reasoning Topology.
Claims we use
- Tree topology — branch, evaluate, backtrack over “thought” units.
- Schedule separation — topology (tree) vs search procedure (BFS, DFS, …).
Author (primary)
Shunyu Yao — full author list in the original paper.
Corpus stance
Context — in-model prompting framework; event capture analogue: ThoughtGraphEmitted with topology: tree (Thinking by Writing).