← 返回论文库

PAL: Program-Aided Language Models

Gao, Madaan, et al. (CMU) · 2022

L3.1 · LLM Agent Patterns & FrameworksICML 2023#reasoning#code

CORE IDEA

用代码作为 reasoning trace：LLM 写 Python solving step，Python 执行得答案。

CONCRETE EXAMPLE

比 CoT 数值准确性显著高。

L-ANCHOR · 为什么在这一层重要

code-as-reasoning

arXiv:2211.10435 ↗

相关论文

Chain-of-Thought Prompting Elicits Reasoning in Large Language Models

Self-Consistency Improves Chain of Thought Reasoning in Language Models

Tree of Thoughts: Deliberate Problem Solving with Large Language Models

Graph of Thoughts: Solving Elaborate Problems with Large Language Models