From 3c375604e1b63b63353b4bc4892b96dcd5316df6 Mon Sep 17 00:00:00 2001 From: "Jieyi Long (THETA Network)" Date: Fri, 9 Jun 2023 03:21:07 -0700 Subject: [PATCH 1/2] Update tot.en.mdx --- pages/techniques/tot.en.mdx | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/pages/techniques/tot.en.mdx b/pages/techniques/tot.en.mdx index a565495..fa9b1d3 100644 --- a/pages/techniques/tot.en.mdx +++ b/pages/techniques/tot.en.mdx @@ -6,7 +6,7 @@ import TOT from '../../img/TOT.png' import TOT2 from '../../img/TOT2.png' import TOT3 from '../../img/TOT3.png' -For complex tasks that require exploration or strategic lookahead, traditional or simple prompting techniques fall short. [Yao et el. (2023)](https://arxiv.org/abs/2305.10601) recently proposed Tree of Thoughts (ToT), a framework that generalizes over chain-of-thought prompting and encourages exploration over thoughts that serve as intermediate steps for general problem solving with language models. +For complex tasks that require exploration or strategic lookahead, traditional or simple prompting techniques fall short. [Yao et el. (2023)](https://arxiv.org/abs/2305.10601) and [Long (2023)](https://arxiv.org/abs/2305.08291) recently proposed Tree of Thoughts (ToT), a framework that generalizes over chain-of-thought prompting and encourages exploration over thoughts that serve as intermediate steps for general problem solving with language models. ToT maintains a tree of thoughts, where thoughts represent coherent language sequences that serve as intermediate steps toward solving a problem. This approach enables an LM to self-evaluate the progress intermediate thoughts make towards solving a problem through a deliberate reasoning process. The LM ability to generate and evaluate thoughts is then combined with search algorithms (e.g., breadth-first search and depth-first search) to enable systematic exploration of thoughts with lookahead and backtracking. @@ -27,4 +27,4 @@ From the results reported in the figure below, ToT substantially outperforms the Image Source: [Yao et el. (2023)](https://arxiv.org/abs/2305.10601) -Code available [here](https://github.com/princeton-nlp/tree-of-thought-llm) \ No newline at end of file +Code available [here](https://github.com/princeton-nlp/tree-of-thought-llm) and [here](https://github.com/jieyilong/tree-of-thought-puzzle-solver) From 7932c65d1ee2764088ef3befce795d833e393bb0 Mon Sep 17 00:00:00 2001 From: "Jieyi Long (THETA Network)" Date: Wed, 14 Jun 2023 17:58:42 -0700 Subject: [PATCH 2/2] Update tot.en.mdx --- pages/techniques/tot.en.mdx | 3 +++ 1 file changed, 3 insertions(+) diff --git a/pages/techniques/tot.en.mdx b/pages/techniques/tot.en.mdx index fa9b1d3..67e7392 100644 --- a/pages/techniques/tot.en.mdx +++ b/pages/techniques/tot.en.mdx @@ -28,3 +28,6 @@ From the results reported in the figure below, ToT substantially outperforms the Image Source: [Yao et el. (2023)](https://arxiv.org/abs/2305.10601) Code available [here](https://github.com/princeton-nlp/tree-of-thought-llm) and [here](https://github.com/jieyilong/tree-of-thought-puzzle-solver) + +At a high level, the main ideas of [Yao et el. (2023)](https://arxiv.org/abs/2305.10601) and [Long (2023)](https://arxiv.org/abs/2305.08291) are similar. Both enhance LLM's capability for complex problem solving through tree search via a multi-round conversation. One of the main difference is that [Yao et el. (2023)](https://arxiv.org/abs/2305.10601) leverages DFS/BFS/beam search, while the tree search strategy (i.e. when to backtrack and backtracking by how many levels, etc.) proposed in [Long (2023)](https://arxiv.org/abs/2305.08291) is driven by a "ToT Controller" trained through reinforcement learning. DFS/BFS/Beam search are generic solution search strategies with no adaptation to specific problems. In comparison, a ToT Controller trained through RL might be able learn from new data set or through self-play (AlphaGo vs brute force search), and hence the RL-based ToT system can continue to evolve and learn new knowledge even with a fixed LLM. +