IterSelectTune: An Iterative Training Framework for Efficient Instruction-Tuning Data Selection

TL;DR

IterSelectTune is an innovative iterative training framework that efficiently selects high-quality instruction data for large language models, reducing human effort and computational costs while improving performance.

Contribution

The paper introduces IterSelectTune, a novel data selection method that outperforms full dataset fine-tuning with only 20% of data and minimal human involvement.

Findings

Outperforms models trained on full datasets across multiple benchmarks.

Reduces instruction tuning data requirements by 80%.

Requires minimal reliance on GPT-4 during data selection.

Abstract

As large language models (LLMs) continue to advance, instruction tuning has become critical for improving their ability to generate accurate and contextually appropriate responses. Although numerous instruction-tuning datasets have been developed to enhance LLM performance, selecting high-quality instruction data from large source datasets typically demands significant human effort. In this work, we introduce $\textbf{IterSelectTune}$, an efficient, cost-effective iterative training policy for selecting high-quality instruction data with no human involvement and limited reliance on GPT-4. By fine-tuning on approximately 20\% of the source data, our method consistently outperforms models fine-tuned on the full dataset across multiple benchmarks and public test datasets. These results highlight the effectiveness of our approach in enhancing LLM performance while reducing the computational resources required for instruction tuning.