Related papers: Long Context is Not Long at All: A Prospector of Long-Dependency Data for Large Language Models

Long Context is Not Long at All: A Prospector of Long-Dependency Data for Large Language Models

URL: http://arxiv.org/abs/2405.17915v1
Date: Tue, 28 May 2024 07:36:56 GMT
Title: Long Context is Not Long at All: A Prospector of Long-Dependency Data for Large Language Models
Authors: Longze Chen, Ziqiang Liu, Wanwei He, Yunshui Li, Run Luo, Min Yang,
Abstract summary: Long-context modeling capabilities are important for large language models (LLMs) in various applications. We propose a data mining framework textbfProLong that can assign each training sample with a long dependency score. Comprehensive experiments on multiple benchmarks indicate that ProLong effectively identifies documents that carry long dependencies.
Score: 13.091271774417867
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Abstract: Long-context modeling capabilities are important for large language models (LLMs) in various applications. However, directly training LLMs with long context windows is insufficient to enhance this capability since some training samples do not exhibit strong semantic dependencies across long contexts. In this study, we propose a data mining framework \textbf{ProLong} that can assign each training sample with a long dependency score, which can be used to rank and filter samples that are more advantageous for enhancing long-context modeling abilities in LLM training. Specifically, we first use delta perplexity scores to measure the \textit{Dependency Strength} between text segments in a given document. Then we refine this metric based on the \textit{Dependency Distance} of these segments to incorporate spatial relationships across long-contexts. Final results are calibrated with a \textit{Dependency Specificity} metric to prevent trivial dependencies introduced by repetitive patterns. Moreover, a random sampling approach is proposed to optimize the computational efficiency of ProLong. Comprehensive experiments on multiple benchmarks indicate that ProLong effectively identifies documents that carry long dependencies and LLMs trained on these documents exhibit significantly enhanced long-context modeling capabilities.