2 papers
cs.CV2026
OnceSelect: Reusable Data Selection for Efficient Multimodal Instruction Tuning
Mingkang Dong, Hongyi Cai, Xiwen Lei +8
Multimodal instruction tuning is widely used to adapt multimodal large language models (MLLMs), yet the large-scale image-text datasets it relies on are often highly redundant. Exi…
cs.CL2025
Low-Confidence Gold: Refining Low-Confidence Samples for Efficient Instruction Tuning
Hongyi Cai, Jie Li, Mohammad Mahdinur Rahman +1
The effectiveness of instruction fine-tuning for Large Language Models is fundamentally constrained by the quality and efficiency of training datasets. This work introduces Low-Con…