Can I read Take the Bull by the Horns: Hard Sample-Reweighted Continual Training Improves LLM Generalization on EtoBox?

Take the Bull by the Horns: Hard Sample-Reweighted Continual Training Improves LLM Generalization by Chen, Xuxi; Wang, Zhendong; Sow, Daouda; Yang, Junjie; Chen, Tianlong; Liang, Yingbin; Zhou, Mingyuan; Wang, Zhangyang is a scholarly article available to read on EtoBox.

What is Take the Bull by the Horns: Hard Sample-Reweighted Continual Training Improves LLM Generalization about?

In the rapidly advancing arena of large language models (LLMs), a key challenge is to enhance their capabilities amid a looming shortage of high-quality training data. Our study starts from an empirical strategy for the light continual training of LLMs using their original pre-training data sets, with a specific focus on selective retention of samples that incur moderately high losses. These samples are deemed informative and beneficial for model refinement, contrasting with the highest-loss samples, which would be discarded due to their correlation with data noise and complexity. We then formalize this strategy into a principled framework of Instance-Reweighted Distributionally Robust Optimization (IR-DRO). IR-DRO is designed to dynamically prioritize the training focus on informative samples through an instance reweighting mechanism, streamlined by a closed-form solution for straightforward integration into established training protocols. Through rigorous experimentation with various models and datasets, our findings indicate that our sample-targeted methods significantly improve LLM performance across multiple benchmarks, in both continual pre-training and instruction tuning sce

Author

Chen, Xuxi; Wang, Zhendong; Sow, Daouda; Yang, Junjie; Chen, Tianlong; Liang, Yingbin; Zhou, Mingyuan; Wang, Zhangyang

Published

2024

Language

EN