0

Self-Taught Self-Correction for Small Language Models

The Self-Taught Self-Correction (STaSC) algorithm improves the performance of small language models through iterative self-correction using self-generated data.

Year
2025
Venue
arXiv 2025
Authors
3
Hosting
Abstract onlyARXIV-DEFAULT

Cite

Notes

Only stored in your browser.

Attribution

Abstract & full text
arxiv.org/abs/2503.08681ARXIV-DEFAULT
TL;DR
Semantic Scholar
Attribution policy →

Abstract

Although large language models (LLMs) have achieved remarkable performance across various tasks, they remain prone to errors. A key challenge is enabling them to self-correct. While prior research has relied on external tools or large proprietary models, this work explores self-correction in small language models (SLMs) through iterative fine-tuning using solely self-generated data. We introduce the Self-Taught Self-Correction (STaSC) algorithm, which incorporates multiple algorithmic design choices. Experimental results on a question-answering task demonstrate that STaSC effectively learns self-correction, leading to significant performance improvements. Our analysis further provides insights into the mechanisms of self-correction and the impact of different design choices on learning dynamics and overall performance. To support future research, we release our user-friendly codebase and lightweight models.

Authors

3