paper-with-me

홈 › Papers

Instruction Matters: A Simple yet Effective Task Selection for Optimized Instruction Tuning of Specific Tasks

2024-04-25 · Changho Lee, Janghoon Han, Seonghyeon Ye, Stanley Jungkyu Choi, Honglak Lee, Kyunghoon Bae

Instruction tuning has been proven effective in enhancing zero-shot generalization across various tasks and in improving the performance of specific tasks. For task-specific improvements, strategically selecting and training on related tasks that provide meaningful supervision is crucial, as this approach enhances efficiency and prevents performance degradation from learning irrelevant tasks. In this light, we introduce a simple yet effective task selection method that leverages instruction information alone to identify relevant tasks, optimizing instruction tuning for specific tasks. Our method is significantly more efficient than traditional approaches, which require complex measurements of pairwise transferability between tasks or the creation of data samples for the target task. Additionally, by aligning the model with the unique instructional template style of the meta-dataset, we enhance its ability to granularly discern relevant tasks, leading to improved overall performance. Experimental results demonstrate that training on a small set of tasks, chosen solely based on the instructions, results in substantial improvements in performance on benchmarks such as P3, Big-Bench, NIV2, and Big-Bench Hard. Significantly, these improvements surpass those achieved by prior task selection methods, highlighting the superiority of our approach.

📄 PDF Abstract BibTeX arXiv:2404.16418

Code (1)

chlee0801/insta 공식 구현 pytorch

Tasks

Zero-shot Generalization

Methods 이 논문이 사용한 방법론

SET Dynamic Sparse Training method where weight mask is updated randomly periodically

Similar Papers 제목 키워드 기반

A Critical Look at Targeted Instruction Selection: Disentangling What Matters (and What Doesn't)

2026-02-16 · Nihal V. Nayak, Paula Rodriguez-Diaz, Neha Hulkund, Sara Beery 외 arxiv

Instruction fine-tuning of large language models (LLMs) often involves selecting a subset of instruction training data from a large candidate pool, using a small query set from the target task. Despite growing interest, …

Data Diversity Matters for Robust Instruction Tuning

2023-11-21 · Alexander Bukharin, Shiyang Li, Zhengyang Wang, Jingfeng Yang 외

Recent works have shown that by curating high quality and diverse instruction tuning datasets, we can significantly improve instruction-following capabilities. However, creating such datasets is difficult and most works …

DiversityInstruction Following

Task--Specificity Score: Measuring How Much Instructions Really Matter for Supervision

2026-02-03 · Pritam Kadasi, Abhishek Upperwal, Mayank Singh arxiv

Instruction tuning is now the default way to train and adapt large language models, but many instruction--input--output pairs are only weakly specified: for a given input, the same output can remain plausible under sever…

From Instance Selection to Fixed-Pool Data Recipe Search for Supervised Fine-Tuning

2026-05-13 · Haodong Wu, Jiahao Zhang, Lijie Hu, Yongqi Zhang arxiv

Supervised fine-tuning (SFT) data selection is commonly formulated as instance ranking: score each example and retain a top-$k$ subset. However, effective SFT training subsets are often produced through ordered curation …

Learning Multi-Indicator Weights for Data Selection: A Joint Task-Model Adaptation Framework with Efficient Proxies

2026-05-10 · Jingze Song, Zihao Chen, Wenqing Chen, Zibin Zheng arxiv

Data selection is a key component of efficient instruction tuning for large language models, as recent work has shown that data quality often matters more than data quantity. Accordingly, prior studies have introduced va…