Reprompting: Automated Chain-of-Thought Prompt Inference Through Gibbs Sampling
This addresses the challenge of manual prompt engineering for large language models, offering a scalable solution for improving reasoning performance in AI applications.
The paper tackles the problem of automatically generating Chain-of-Thought prompts for reasoning tasks without human intervention, achieving an average improvement of +9.4 points over human-written prompts across 20 tasks.
We introduce Reprompting, an iterative sampling algorithm that automatically learns the Chain-of-Thought (CoT) recipes for a given task without human intervention. Through Gibbs sampling, Reprompting infers the CoT recipes that work consistently well for a set of training samples by iteratively sampling new recipes using previously sampled recipes as parent prompts to solve other training problems. We conduct extensive experiments on 20 challenging reasoning tasks. Results show that Reprompting outperforms human-written CoT prompts substantially by +9.4 points on average. It also achieves consistently better performance than the state-of-the-art prompt optimization and decoding algorithms.