ACL Findings2024
SMART: Submodular Data Mixture Strategy for Instruction Tuning
H. S. V. N. S. Kowndinya Renduchintala, Sumit Bhatia, Ganesh Ramakrishnan
Abstract
Instruction Tuning involves finetuning a language model on a collection of instructionformatted datasets in order to enhance the generalizability of the model to unseen tasks. Studies have shown the importance of balancing different task proportions during finetuning, but finding the right balance remains challenging. Unfortunately, there's currently no systematic method beyond manual tuning or relying on practitioners' intuition. In this paper, we introduce SMART (Submodular data Mixture strAtegy for instRuction Tuning)a novel data mixture strategy which makes use of a submodular function to assign importance scores to tasks which are then used to determine the mixture weights. Given a fine-tuning budget, SMART redistributes the budget among tasks and selects non-redundant samples from each task. Experimental results demonstrate that SMART significantly outperforms traditional methods such as examples proportional mixing and equal mixing. Furthermore, SMART facilitates the creation of data mixtures based on a few representative subsets of tasks alone and through task pruning analysis, we reveal that in a limited budget setting, allocating budget among a subset of representative tasks yields superior performance compared to distributing the budget among all tasks. The code for reproducing our results is open-sourced at https://github.com/kowndinya- renduchintala/SMART.
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
8 top-tier papers have cited it since. Know what they found before your reviewers do.
Lune can draw on
the full paper · the top-tier papers that cite it · related work
Free to start. No credit card required.
Lineage
- PASER: Post-Training Data Selection for Efficient Pruned Large Language Model Recovery cites Measuring Massive Multitask Language Understanding
- PASER: Post-Training Data Selection for Efficient Pruned Large Language Model Recovery cites Finetuned Language Models are Zero-Shot Learners
- Minibatch selection for Language Models via Partition Matroid Constrained Gradient Matching cites Measuring Massive Multitask Language Understanding
- Measuring Massive Multitask Language Understanding cites Language Models are Few-Shot Learners
- Submodular Optimization for Minimal Augmentation in Robust Language Model Alignment cites Measuring Massive Multitask Language Understanding
- Neural Networks for Learnable and Scalable Influence Estimation of Instruction Fine-Tuning Data cites Measuring Massive Multitask Language Understanding
- Large Language Models are Zero-Shot Reasoners cites Language Models are Few-Shot Learners
- Diversity as a Reward: Fine-Tuning LLMs on a Mixture of Domain-Undetermined Data cites Measuring Massive Multitask Language Understanding
- Finetuned Language Models are Zero-Shot Learners cites Language Models are Few-Shot Learners
Cited by top-tier papers8
- Diversity as a Reward: Fine-Tuning LLMs on a Mixture of Domain-Undetermined DataZhenqing Ling, Daoyuan Chen, Liuyi Yao, Qianli Shen et al.NeurIPS 2025 · 14 citations
- PASER: Post-Training Data Selection for Efficient Pruned Large Language Model RecoveryBowei He, Lihao Yin, Huiling Zhen, Xiaokun Zhang et al.ICLR 2026 · 5 citations
- Neural Networks for Learnable and Scalable Influence Estimation of Instruction Fine-Tuning DataIshika Agarwal, Dilek Hakkani-TurNeurIPS 2025 · 2 citations
- Minibatch selection for Language Models via Partition Matroid Constrained Gradient MatchingPrayas Agrawal, Prateek Chanda, Ishita Khatri, Ganesh Ramakrishnan et al.ICML 2026
- Submodular Optimization for Minimal Augmentation in Robust Language Model AlignmentCHING-CHIA KAO, Chia-Mu Yu, Chun-Shien Lu, Chu-song ChenICML 2026
Builds on31
- Language Models are Few-Shot LearnersTom B. Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah et al.NeurIPS 2020 · 64,255 citations
- Large Language Models are Zero-Shot ReasonersTakeshi Kojima, Shixiang Shane Gu, Machel Reid, Yutaka Matsuo et al.NeurIPS 2022 · 8,168 citations
- Measuring Massive Multitask Language UnderstandingDan Hendrycks, Collin Burns, Steven Basart, Andy Zou et al.ICLR 2021 · 7,905 citations
- Self-Refine: Iterative Refinement with Self-FeedbackAman Madaan, Niket Tandon, Prakhar Gupta, Skyler Hallinan et al.NeurIPS 2023 · 4,972 citations
- Finetuned Language Models are Zero-Shot LearnersJason Wei, Maarten Bosma, Vincent Y. Zhao, Kelvin Guu et al.ICLR 2022 · 4,966 citations
Related papers
- Sweeping Heterogeneity with Smart MoPs: Mixture of Prompts for LLM Task AdaptationChen Dun, Mirian del Carmen Hipolito Garcia, Guoqing Zheng, Ahmed Hassan Awadallah et al.AAAI 2025 · 7 citations
- DynamixSFT: Dynamic Mixture Optimization of Instruction Tuning CollectionsHaebin Shin, Lei Ji, Xiao Liu, Zhiwei Yu et al.ACL Findings 2026
- Mastering Collaborative Multi-Modal Data Selection: A Focus on Informativeness, Uniqueness, and RepresentativenessQifan Yu, Zhebei Shen, Zhongqi Yue, Yang Wu et al.ICCV 2025 · 1 citation
- Learning Instructions with Unlabeled Data for Zero-Shot Cross-Task GeneralizationYuxian Gu, Pei Ke, Xiaoyan Zhu, Minlie HuangEMNLP 2022 · 3 citations
- MIG: Automatic Data Selection for Instruction Tuning by Maximizing Information Gain in Semantic SpaceYicheng Chen, Yining Li, Kai Hu, Zerun Ma et al.ACL Findings 2025 · 2 citations
