QLESS: A Quantized Approach for Data Valuation and Selection in Large Language Model Fine-Tuning

TL;DR

QLESS introduces a memory-efficient method combining gradient quantization and low-rank approximation for data valuation in large language model fine-tuning, maintaining performance with significantly reduced memory requirements.

Contribution

It presents a novel quantized gradient similarity search framework that enables scalable data selection for LLM fine-tuning under strict memory constraints.

Findings

QLESS achieves comparable data selection performance to LESS.

Memory usage is reduced by up to 16x with QLESS.

1-bit gradient quantization still preserves data valuation quality.

Abstract

Fine-tuning large language models (LLMs) is often constrained by the computational costs of processing massive datasets. We propose \textbf{QLESS} (Quantized Low-rank Gradient Similarity Search), which integrates gradient quantization with the LESS framework to enable memory-efficient data valuation and selection. QLESS employs a two-step compression process: first, it obtains low-dimensional gradient representations through LoRA-based random projection; then, it quantizes these gradients to low-bitwidth representations. Experiments on multiple LLM architectures (LLaMA, Mistral, Qwen) and benchmarks (MMLU, BBH, TyDiQA) show that QLESS achieves comparable data selection performance to LESS while reducing memory usage by up to 16x. Even 1-bit gradient quantization preserves data valuation quality. These findings underscore QLESS as a practical, scalable approach to identifying informative examples within strict memory constraints.