Why Can Large Language Models Generate Correct Chain-of-thoughts?

Tutunov Rasul, Grosnit Antoine, Ziomek Juliusz, Wang Jun, Bou-ammar Haitham. Arxiv 2023

[Paper]    
Prompting Tools

This paper delves into the capabilities of large language models (LLMs), specifically focusing on advancing the theoretical comprehension of chain-of-thought prompting. We investigate how LLMs can be effectively induced to generate a coherent chain of thoughts. To achieve this, we introduce a two-level hierarchical graphical model tailored for natural language generation. Within this framework, we establish a compelling geometrical convergence rate that gauges the likelihood of an LLM-generated chain of thoughts compared to those originating from the true language. Our findings provide a theoretical justification for the ability of LLMs to produce the correct sequence of thoughts (potentially) explaining performance gains in tasks demanding reasoning skills.

Similar Work