paper-with-me

Papers

Diversity Measurement and Subset Selection for Instruction Tuning Datasets

2024-02-04 · Peiqi Wang, Yikang Shen, Zhen Guo, Matthew Stallone, Yoon Kim, Polina Golland, Rameswar Panda

We aim to select data subsets for the fine-tuning of large language models to more effectively follow instructions. Prior work has emphasized the importance of diversity in dataset curation but relied on heuristics such as the number of tasks. In this paper, we use determinantal point processes to capture the diversity and quality of instruction tuning datasets for subset selection. We propose to measure dataset diversity with log determinant distance that is the distance between the dataset of interest and a maximally diverse reference dataset. Our experiments demonstrate that the proposed diversity measure in the normalized weight gradient space is correlated with downstream instruction-following performance. Consequently, it can be used to inform when data selection is the most helpful and to analyze dataset curation strategies. We demonstrate the utility of our approach on various instruction tuning datasets.

📄 PDF Abstract BibTeX arXiv:2402.02318

Code (0)

등록된 구현이 없습니다.

Tasks

DiversityInstruction FollowingPoint Processes

Similar Papers 제목 키워드 기반

CoIDO: Efficient Data Selection for Visual Instruction Tuning via Coupled Importance-Diversity Optimization

2025-10-11 · Yichen Yan, Ming Zhong, Qi Zhu, Xiaoling Gu 외 arxiv

Multimodal large language models (MLLMs) rely heavily on instruction tuning to align vision and language capabilities, yet the computational cost of training on large-scale datasets remains a major bottleneck. Existing d…

D3: Diversity, Difficulty, and Dependability-Aware Data Selection for Sample-Efficient LLM Instruction Tuning

2025-03-14 · Jia Zhang, Chen-Xi Zhang, Yao Liu, Yi-Xuan Jin 외

Recent advancements in instruction tuning for large language models (LLMs) suggest that a small, high-quality dataset can significantly equip LLMs with instruction-following capabilities, outperforming large datasets oft…

DiversityInstruction Following

InsBank: Evolving Instruction Subset for Ongoing Alignment

2025-02-17 · Jiayi Shi, Yiwei Li, Shaoxiong Feng, Peiwen Yuan 외

Large language models (LLMs) typically undergo instruction tuning to enhance alignment. Recent studies emphasize that quality and diversity of instruction data are more crucial than quantity, highlighting the need to sel…

Diversity

Diversify and Conquer: Diversity-Centric Data Selection with Iterative Refinement

2024-09-17 · Simon Yu, Liangyu Chen, Sara Ahmadian, Marzieh Fadaee

Finetuning large language models on instruction data is crucial for enhancing pre-trained knowledge and improving instruction-following capabilities. As instruction datasets proliferate, selecting optimal data for effect…

Active LearningDiversityInstruction FollowingMath+1

MIG: Automatic Data Selection for Instruction Tuning by Maximizing Information Gain in Semantic Space

2025-04-18 · Yicheng Chen, Yining Li, Kai Hu, Zerun Ma 외

Data quality and diversity are key to the construction of effective instruction-tuning datasets. % With the increasing availability of open-source instruction-tuning datasets, it is advantageous to automatically select h…

Diversity