AI RESEARCH
TRIM: Token-wise Attention-Derived Saliency for Data-Efficient Instruction Tuning
arXiv CS.LG
•
ArXi:2510.07118v3 Announce Type: replace-cross Instruction tuning is essential for aligning large language models (LLMs) to downstream tasks and commonly relies on large, diverse corpora. However, small, high-quality subsets, known as coresets, can deliver comparable or superior results, though curating them remains challenging. Existing methods often rely on coarse, sample-level signals like gradients, an approach that is computationally expensive and overlooks fine-grained features. To address this, we