METEOR: Evolutionary Journey of Large Language Models from Guidance to Self-Growth

TL;DR

This paper introduces the Meteor method, a three-phase training approach that enables large language models to autonomously evolve and improve their domain expertise through self-guided refinement, leading to significant performance gains.

Contribution

The Meteor method provides a unified, effective framework for guiding the self-evolution of large language models across multiple training phases.

Findings

Significant improvements in accuracy and relevance across domain-specific tasks.

Enhanced model coherence and reliability through self-evolution strategies.

Effective autonomous refinement of domain knowledge.

Abstract

Model evolution enables learning from feedback to refine experiences and update skills, transforming models from having no domain knowledge to becoming domain experts. However, there is currently no unified and effective method for guiding this evolutionary process. To address this gap, we propose the Meteor method, which includes three training phases: weak-to-strong data distillation, iterative training, and self-evolution strategies. Each phase maximizes the model's inherent domain capabilities, allowing it to autonomously refine its domain knowledge and enhance performance. Experiments demonstrate that our approach significantly improves accuracy, completeness, relevance, coherence, and reliability across domain-specific tasks.