---
title: "DynamixSFT — paper by Shin et al. (2026) — summary and ideas"
description: "Instruction-tuning collections such as TÜLU combine many heterogeneous datasets, yet their mixing ratios are usually fixed by… 53 references."
authors: ["Haebin Shin", "Lei Ji", "Xiao Liu", "Zhiwei Yu", "Hyunwoo Yoo", "Qi Chen", "Yeyun Gong"]
year: 2026
venue: "Findings of the Association for Computational Linguistics: ACL 2026"
source: "https://aclanthology.org/2026.findings-acl.1972/"
doi: "10.18653/v1/2026.findings-acl.1972"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/shin-2026-dynamixsft-dynamic-mixture-optimization-instruction/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/shin-2026-dynamixsft-dynamic-mixture-optimization-instruction/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/28/) / DynamixSFT: Dynamic Mixture Optimization of Instruction Tuning Collections

# DynamixSFT: Dynamic Mixture Optimization of Instruction Tuning Collections

**Authors:** Haebin Shin, [Lei Ji](https://md.research.elamur.ai/free/en/authors/lei-ji/), [Xiao Liu](https://md.research.elamur.ai/free/en/authors/xiao-liu/), [Zhiwei Yu](https://md.research.elamur.ai/free/en/authors/zhiwei-yu/), Hyunwoo Yoo, [Qi Chen](https://md.research.elamur.ai/free/en/authors/qi-chen/) and [Yeyun Gong](https://md.research.elamur.ai/free/en/authors/yeyun-gong/) · **Year:** 2026 · **Venue:** [Findings of the Association for Computational Linguistics: ACL 2026](https://md.research.elamur.ai/free/en/venues/findings/) ([Findings of the Association for Computational Linguistics 2026](https://md.research.elamur.ai/free/en/venues/findings/2026/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2026.findings-acl.1972/) · [PDF](https://aclanthology.org/2026.findings-acl.1972.pdf) · [DOI 10.18653/v1/2026.findings-acl.1972](https://doi.org/10.18653/v1/2026.findings-acl.1972) · [OpenAlex](https://openalex.org/W4414460334)

**Citations:** 0 citations (OpenAlex, 2026-10-14)

## Summary
Instruction-tuning collections such as TÜLU combine many heterogeneous datasets, yet their mixing ratios are usually fixed by hand or need extra proxy models and validation sets to adjust. DynamixSFT addresses this by automatically re-weighting datasets during supervised fine-tuning. Each dataset is treated as an arm in a multi-armed bandit. Sampling probabilities come from a Prior-scaled Boltzmann Exploration, which multiplies the usual softmax over rewards by the original dataset proportions. This keeps the evolving mixture close to the collection's intended coverage while still letting it adapt. A minimum floor probability ensures no dataset is starved. Rewards come from a 1-Step Look-ahead Reward: at fixed intervals, the method estimates how much a short trial update on each dataset would lower the model's loss, then smooths the result with an EMA. The approach was tested on the TÜLU-2-mixture and TÜLU-3-mixture with LLaMA models across 10 benchmarks. It beats both static and dynamic baselines at modest extra cost, offering a practical, supervision-free way to use large SFT collections.

## Key points
- Collections: TÜLU-2-mixture (16 datasets) and TÜLU-3-mixture (19 datasets). Models: LLaMA3.2 1B and LLaMA3.1 8B. Evaluation covers 10 benchmarks spanning knowledge, reasoning, math, coding, and instruction following (e.g., MMLU, BBH, GSM8K, MATH, IFEval).
- In a pilot study, TÜLU-2 data was split into high, medium, and low NLL groups using Qwen3-30B-A3B. High-loss data favored knowledge and reasoning, while low-loss data favored coding and math. Changing the training order also shifted results, which motivated state-dependent mixing.
- Compared with naive sampling, overhead is only about 12.7% (K=19, update every 50 steps). By contrast, HBO adds +139%, MultiUAT +380%, and MultiDDS +760%.
- Removing the prior term causes a substantial performance drop, with the mixture staying near uniform. Starting from proportional initialization without the prior still underperforms the full method.
- Main limitations: weights are assigned per dataset rather than per instance, and there is no fine-grained per-domain analysis.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

## Institutions

- [Microsoft](https://md.research.elamur.ai/free/en/companies/microsoft/) — Company, Redmond, United States · 325 works · 142,931 citations
- [Beijing Academy of Artificial Intelligence](https://md.research.elamur.ai/free/en/labs/beijing-academy-of-artificial-intelligence/) — Research institute, Beijing, China · 32 works · 7,511 citations
- [Drexel University](https://md.research.elamur.ai/free/en/universities/drexel-university/) — University, Philadelphia, United States · 4 works · 1,443 citations

📚 Shelves (6): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 16,667 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 16,194 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 16,039 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 15,891 works · [🎯 Decision Sciences](https://md.research.elamur.ai/free/en/shelves/decision-sciences/) · 209 works · [🎯 Management Science and Operations Research](https://md.research.elamur.ai/free/en/shelves/decision-sciences/management-science-and-operations-research/) · 108 works

🧭 Topics (1): [🧠 Language models & pretraining](https://md.research.elamur.ai/free/en/topics/language-models/) · 3,320 works

## Cite

Cite as: Haebin Shin, Lei Ji, Xiao Liu, Zhiwei Yu, Hyunwoo Yoo, Qi Chen and Yeyun Gong. 2026. DynamixSFT: Dynamic Mixture Optimization of Instruction Tuning Collections. Findings of the Association for Computational Linguistics: ACL 2026.

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@inproceedings{shin2026dynamixsft,
  title = {{DynamixSFT: Dynamic Mixture Optimization of Instruction Tuning Collections}},
  author = {Haebin Shin, Lei Ji, Xiao Liu, Zhiwei Yu, Hyunwoo Yoo, Qi Chen and Yeyun Gong},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.1972},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.1972}
}
```

- [← previous Dynamic Tool Dependency Retrieval for Lightweight Function Calling](https://md.research.elamur.ai/free/en/papers/patel-2026-dynamic-tool-dependency-retrieval-lightweight/)
- [next → D²Plan: Dual-Agent Dynamic Global Planning for Complex Retrieval-Augmented Reasoning](https://md.research.elamur.ai/free/en/papers/luo-2026-d-2-plan-dual-agent/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/28/)

## References

53 references, 30 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Alon Albalak, Liangming Pan, Colin Raffel, and William Yang Wang. 2023. [Efficient online data mixing for language model pre-training](https://md.research.elamur.ai/free/en/papers/albalak-2023-efficient-online-data-mixing-language/). Preprint, arXiv:2312.02406.
2. Faeze Brahman, Sachin Kumar, Vidhisha Balachandran, Pradeep Dasigi, Valentina Pyatkin, Abhilasha Ravichander, Sarah Wiegreffe, Nouha Dziri, Khyathi Chandu, Jack Hessel, Yulia Tsvetkov, Noah A. Smith, Yejin Choi, and Hannaneh Hajishirzi. 2024. [The art of saying no: Contextual noncompliance in languag…](https://md.research.elamur.ai/free/en/papers/brahman-2024-art-saying-no-contextual-noncompliance/)
3. Isabel Cachola, Kyle Lo, Arman Cohan, and Daniel Weld. 2020. [TLDR: Extreme summarization of scientific documents](https://md.research.elamur.ai/free/en/papers/cachola-2020-tldr-extreme-summarization-scientific-documents/). In Findings of EMNLP.
4. Nicolò Cesa-Bianchi, Claudio Gentile, Gabor Lugosi, and Gergely Neu. 2017. Boltzmann exploration done right. In NeurIPS. [find](https://scholar.google.com/scholar?q=Nicol%C3%B2+Cesa%2DBianchi%2C+Claudio+Gentile%2C+Gabor+Lugosi%2C+and+Gergely+Neu%2E+2017%2E+Boltzmann+exploration+done+right%2E+In+NeurIPS%2E)
5. Sahil Chaudhary. 2023. Code alpaca: An instruction-following llama model for code generation. [https://github.com/sahil280114/codealpaca](https://github.com/sahil280114/codealpaca). [↗](https://github.com/sahil280114/codealpaca)
6. Mark Chen, Jerry Tworek, Heewoo Jun, Qiming Yuan, Henrique Ponde de Oliveira Pinto, Jared Kaplan, Harri Edwards, Yuri Burda, Nicholas Joseph, Greg Brockman, Alex Ray, Raul Puri, Gretchen Krueger, Michael Petrov, Heidy Khlaaf, Girish Sastry, Pamela Mishkin, Brooke Chan, Scott Gray, and 39 others. 202… — [Evaluating Large Language Models Trained on Code](https://md.research.elamur.ai/free/en/papers/chen-2021-evaluating-large-language-models-trained/)
7. Xiaoyin Chen, Jiarui Lu, Minsu Kim, Dinghuai Zhang, Jian Tang, Alexandre Piché, Nicolas Gontier, Yoshua Bengio, and Ehsan Kamalloo. 2025. Self-evolving curriculum for llm reasoning. Preprint, arXiv:2505.14970. [find](https://scholar.google.com/scholar?q=Xiaoyin+Chen%2C+Jiarui+Lu%2C+Minsu+Kim%2C+Dinghuai+Zhang%2C+Jian+Tang%2C+Alexandre+Pich%C3%A9%2C+Nicolas+Gontier%2C+Yoshua+Bengio%2C+and+Ehsan+Kamalloo%2E+2025%2E+Self%2Devolving+curriculum+for+llm+reasoning%2E+Preprint%2C+arXiv%3A25)
8. Karl Cobbe, Vineet Kosaraju, Mohammad Bavarian, Mark Chen, Heewoo Jun, Lukasz Kaiser, Matthias Plappert, Jerry Tworek, Jacob Hilton, Reiichiro Nakano, Christopher Hesse, and John Schulman. 2021. [Training verifiers to solve math word problems](https://md.research.elamur.ai/free/en/papers/cobbe-2021-training-verifiers-solve-math-word/). Preprint, arXiv:2110.14168.
9. Pradeep Dasigi, Kyle Lo, Iz Beltagy, Arman Cohan, Noah A. Smith, and Matt Gardner. 2021. [A dataset of information-seeking questions and answers anchored in research papers](https://md.research.elamur.ai/free/en/papers/dasigi-2021-dataset-information-seeking-questions-answers/). In NAACL.
10. Dheeru Dua, Yizhong Wang, Pradeep Dasigi, Gabriel Stanovsky, Sameer Singh, and Matt Gardner. 2019. [DROP: A reading comprehension benchmark requiring discrete reasoning over paragraphs](https://md.research.elamur.ai/free/en/papers/dua-2019-drop-reading-comprehension-benchmark-requiring/). In NAACL.
11. Simin Fan, Matteo Pagliardini, and Martin Jaggi. 2024a. Doge: Domain reweighting with generalization estimation. Preprint, arXiv:2310.15393. — [arXiv:2310.15393](https://arxiv.org/abs/2310.15393) [find](https://scholar.google.com/scholar?q=Simin+Fan%2C+Matteo+Pagliardini%2C+and+Martin+Jaggi%2E+2024a%2E+Doge%3A+Domain+reweighting+with+generalization+estimation%2E+Preprint%2C+arXiv%3A2310%2E15393%2E+%E2%80%94+%5BarXiv%3A2310%2E15393%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F2310%2E15393%29)
12. Simin Fan, Matteo Pagliardini, and Martin Jaggi. 2024b. DOGE: Domain reweighting with generalization estimation. In ICML. [find](https://scholar.google.com/scholar?q=Simin+Fan%2C+Matteo+Pagliardini%2C+and+Martin+Jaggi%2E+2024b%2E+DOGE%3A+Domain+reweighting+with+generalization+estimation%2E+In+ICML%2E)
13. Aaron Grattafiori, Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Abhishek Kadian, Ahmad Al-Dahle, Aiesha Letman, Akhil Mathur, Alan Schelten, Alex Vaughan, Amy Yang, Angela Fan, Anirudh Goyal, Anthony Hartshorn, Aobo Yang, Archi Mitra, Archie Sravankumar, Artem Korenev, Arthur Hinsvark, and 542 o… — [The Llama 3 Herd of Models](https://md.research.elamur.ai/free/en/papers/grattafiori-2024-llama-3-herd-models/)
14. Harsh Gupta, Seo Taek Kong, R. Srikant, and Weina Wang. 2019. Almost boltzmann exploration. Preprint, arXiv:1901.08708. — [arXiv:1901.08708](https://arxiv.org/abs/1901.08708) [find](https://scholar.google.com/scholar?q=Harsh+Gupta%2C+Seo+Taek+Kong%2C+R%2E+Srikant%2C+and+Weina+Wang%2E+2019%2E+Almost+boltzmann+exploration%2E+Preprint%2C+arXiv%3A1901%2E08708%2E+%E2%80%94+%5BarXiv%3A1901%2E08708%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F1901%2E08708%29)
15. Dan Hendrycks, Collin Burns, Steven Basart, Andy Zou, Mantas Mazeika, Dawn Song, and Jacob Steinhardt. 2021. [Measuring massive multitask language understanding](https://md.research.elamur.ai/free/en/papers/hendrycks-2020-measuring-massive-multitask-language-understanding/). In ICLR.
16. Hamish Ivison, Yizhong Wang, Valentina Pyatkin, Nathan Lambert, Matthew Peters, Pradeep Dasigi, Joel Jang, David Wadden, Noah A. Smith, Iz Beltagy, and Hannaneh Hajishirzi. 2023. [Camels in a changing climate: Enhancing lm adaptation with tulu 2](https://md.research.elamur.ai/free/en/papers/ivison-2023-camels-changing-climate-enhancing-lm/). Preprint, arXiv:2311.10702.
17. Andreas Köpf, Yannic Kilcher, Dimitri von Rütte, Sotiris Anagnostidis, Zhi Rui Tam, Keith Stevens, Abdullah Barhoum, Duc Nguyen, Oliver Stanley, Richárd Nagyfi, Shahul ES, Sameer Suri, David Glushkov, Arnav Dantuluri, Andrew Maguire, Christoph Schuhmann, Huu Nguyen, and Alexander Mattick. 2023. Open… — [OpenAssistant Conversations -- Democratizing Large Language Model Alignment](https://md.research.elamur.ai/free/en/papers/kopf-2023-openassistant-conversations-democratizing-large-language/)
18. Nathan Lambert, Jacob Morrison, Valentina Pyatkin, Shengyi Huang, Hamish Ivison, Faeze Brahman, Lester James V. Miranda, Alisa Liu, Nouha Dziri, Shane Lyu, Yuling Gu, Saumya Malik, Victoria Graf, Jena D. Hwang, Jiangjiang Yang, Ronan Le Bras, Oyvind Tafjord, Chris Wilhelm, Luca Soldaini, and 4 other… — [Tulu 3: Pushing Frontiers in Open Language Model Post-Training](https://md.research.elamur.ai/free/en/papers/lambert-2024-tulu-3-pushing-frontiers-open/)
19. Eric Lehman, Jay DeYoung, Regina Barzilay, and Byron C. Wallace. 2019. [Inferring which medical treatments work from reports of clinical trials](https://md.research.elamur.ai/free/en/papers/lehman-2019-inferring-which-medical-treatments-work/). In NAACL.
20. Jia LI, Edward Beeching, Lewis Tunstall, Ben Lipkin, Roman Soletskyi, Shengyi Costa Huang, Kashif Rasul, Longhui Yu, Albert Jiang, Ziju Shen, Zihan Qin, Bin Dong, Li Zhou, Yann Fleureau, Guillaume Lample, and Stanislas Polu. 2024. Numinamath tir. [https://huggingface.co/AI-MO/](https://huggingface.co/AI-MO/) NuminaMath-TIR. [↗](https://huggingface.co/AI-MO/)
21. Peng Li, Yeye He, Dror Yashar, Weiwei Cui, Song Ge, Haidong Zhang, Danielle Rifinski Fainman, Dongmei Zhang, and Surajit Chaudhuri. 2024. Table-gpt: Table fine-tuned gpt for diverse table tasks. Proceedings of the ACM on Management of Data, 2(3). [find](https://scholar.google.com/scholar?q=Peng+Li%2C+Yeye+He%2C+Dror+Yashar%2C+Weiwei+Cui%2C+Song+Ge%2C+Haidong+Zhang%2C+Danielle+Rifinski+Fainman%2C+Dongmei+Zhang%2C+and+Surajit+Chaudhuri%2E+2024%2E+Table%2Dgpt%3A+Table+fine%2Dtuned+gpt+for+diverse+table+tasks%2E+Proce)
22. Wing Lian, Bleys Goodson, Eugene Pentland, Austin Cook, Chanvichet Vong, and "Teknium". 2023. Openorca: An open dataset of gpt augmented flan reasoning traces. [https://huggingface.co/](https://huggingface.co/) datasets/Open-Orca/OpenOrca. [↗](https://huggingface.co/)
23. Stephanie Lin, Jacob Hilton, and Owain Evans. 2022. [TruthfulQA: Measuring how models mimic human falsehoods](https://md.research.elamur.ai/free/en/papers/lin-2022-truthfulqa-measuring-how-models-mimic/). In ACL.
24. Jiawei Liu, Chunqiu Steven Xia, Yuyao Wang, and LINGMING ZHANG. 2023. [Is your code generated by chatGPT really correct? rigorous evaluation of large language models for code generation](https://md.research.elamur.ai/free/en/papers/liu-2023-your-code-generated-chatgpt-really/). In NeurIPS.
25. Qian Liu, Xiaosen Zheng, Niklas Muennighoff, Guangtao Zeng, Longxu Dou, Tianyu Pang, Jing Jiang, and Min Lin. 2025. [Regmix: Data mixture as regression for language model pre-training](https://md.research.elamur.ai/free/en/papers/liu-2024-regmix-data-mixture-regression-language/). Preprint, arXiv:2407.01492.
26. Shayne Longpre, Le Hou, Tu Vu, Albert Webson, Hyung Won Chung, Yi Tay, Denny Zhou, Quoc V. Le, Barret Zoph, Jason Wei, and Adam Roberts. 2023. [The flan collection: Designing data and methods for effective instruction tuning](https://md.research.elamur.ai/free/en/papers/longpre-2023-flan-collection-designing-data-methods/). Preprint, arXiv:2301.13688.
27. Yi Luan, Luheng He, Mari Ostendorf, and Hannaneh Hajishirzi. 2018. [Multi-task identification of entities, relations, and coreference for scientific knowledge graph construction](https://md.research.elamur.ai/free/en/papers/luan-2018-multi-task-identification-entities-relations/). In EMNLP.
28. Ziyang Luo, Can Xu, Pu Zhao, Qingfeng Sun, Xiubo Geng, Wenxiang Hu, Chongyang Tao, Jing Ma, Qingwei Lin, and Daxin Jiang. 2023. Wizardcoder: Empowering code large language models with evol-instruct. Preprint, arXiv:2306.08568. [find](https://scholar.google.com/scholar?q=Ziyang+Luo%2C+Can+Xu%2C+Pu+Zhao%2C+Qingfeng+Sun%2C+Xiubo+Geng%2C+Wenxiang+Hu%2C+Chongyang+Tao%2C+Jing+Ma%2C+Qingwei+Lin%2C+and+Daxin+Jiang%2E+2023%2E+Wizardcoder%3A+Empowering+code+large+language+models+with+evol%2Dinstruct%2E+P)
29. Alex Mallen, Akari Asai, Victor Zhong, Rajarshi Das, Daniel Khashabi, and Hannaneh Hajishirzi. 2023. When not to trust language models: Investigating effectiveness of parametric and non-parametric memories. In ACL. [find](https://scholar.google.com/scholar?q=Alex+Mallen%2C+Akari+Asai%2C+Victor+Zhong%2C+Rajarshi+Das%2C+Daniel+Khashabi%2C+and+Hannaneh+Hajishirzi%2E+2023%2E+When+not+to+trust+language+models%3A+Investigating+effectiveness+of+parametric+and+non%2Dparametric+mem)
30. Meta. 2024. Llama 3.2 model card. [https://github.com/meta-llama/llama-models/](https://github.com/meta-llama/llama-models/) blob/main/models/llama3_2/MODEL_CARD.md. [↗](https://github.com/meta-llama/llama-models/)
31. Baolin Peng, Chunyuan Li, Pengcheng He, Michel Galley, and Jianfeng Gao. 2023. [Instruction tuning with gpt-4](https://md.research.elamur.ai/free/en/papers/peng-2023-instruction-tuning-gpt-4/). Preprint, arXiv:2304.03277.
32. Nazneen Rajani, Lewis Tunstall, Edward Beeching, Nathan Lambert, Alexander M. Rush, and Thomas Wolf. 2023. No robots. [https://huggingface.co/](https://huggingface.co/) datasets/HuggingFaceH4/no_robots. [↗](https://huggingface.co/)
33. Shivalika Singh, Freddie Vargus, Daniel Dsouza, Börje F. Karlsson, Abinaya Mahendiran, Wei-Yin Ko, Herumb Shandilya, Jay Patel, Deividas Mataciunas, Laura OMahony, Mike Zhang, Ramith Hettiarachchi, Joseph Wilson, Marina Machado, Luisa Souza Moura, Dominik Krzemiński, Hakimeh Fadaei, Irem Ergün, Ife… — [Aya Dataset: An Open-Access Collection for Multilingual Instruction Tuning](https://md.research.elamur.ai/free/en/papers/singh-2024-aya-dataset-open-access-collection/)
34. Richard S Sutton, Andrew G Barto, and 1 others. 1998. Reinforcement learning: An introduction. MIT press Cambridge. [find](https://scholar.google.com/scholar?q=Richard+S+Sutton%2C+Andrew+G+Barto%2C+and+1+others%2E+1998%2E+Reinforcement+learning%3A+An+introduction%2E+MIT+press+Cambridge%2E)
35. Mirac Suzgun, Nathan Scales, Nathanael Schärli, Sebastian Gehrmann, Yi Tay, Hyung Won Chung, Aakanksha Chowdhery, Quoc V. Le, Ed H. Chi, Denny Zhou, and Jason Wei. 2022. [Challenging big-bench tasks and whether chain-of-thought can solve them](https://md.research.elamur.ai/free/en/papers/suzgun-2023-challenging-big-bench-tasks-whether/). Preprint, arXiv:2210.09261.
36. Shubham Toshniwal, Wei Du, Ivan Moshkov, Branislav Kisacanin, Alexan Ayrapetyan, and Igor Gitman. 2024. Openmathinstruct-2: Accelerating ai for math with massive open-source instruction data. Preprint, arXiv:2410.01560. [find](https://scholar.google.com/scholar?q=Shubham+Toshniwal%2C+Wei+Du%2C+Ivan+Moshkov%2C+Branislav+Kisacanin%2C+Alexan+Ayrapetyan%2C+and+Igor+Gitman%2E+2024%2E+Openmathinstruct%2D2%3A+Accelerating+ai+for+math+with+massive+open%2Dsource+instruction+data%2E+Preprint)
37. David Wadden, Shanchuan Lin, Kyle Lo, Lucy Lu Wang, Madeleine van Zuylen, Arman Cohan, and Hannaneh Hajishirzi. 2020. [Fact or fiction: Verifying scientific claims](https://md.research.elamur.ai/free/en/papers/wadden-2020-fact-or-fiction-verifying-scientific/). In EMNLP.
38. David Wadden, Kejian Shi, Jacob Morrison, Alan Li, Aakanksha Naik, Shruti Singh, Nitzan Barzilay, Kyle Lo, Tom Hope, Luca Soldaini, Shannon Zejiang Shen, Doug Downey, Hannaneh Hajishirzi, and Arman Cohan. 2025. [SciRIFF: A resource to enhance language model instruction-following over scientific liter…](https://md.research.elamur.ai/free/en/papers/wadden-2025-sciriff-resource-enhance-language-model/)
39. Renxi Wang, Haonan Li, Minghao Wu, Yuxia Wang, Xudong Han, Chiyu Zhang, and Timothy Baldwin. 2024. Demystifying instruction mixing for fine-tuning large language models. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 4: Student Research Workshop). [find](https://scholar.google.com/scholar?q=Renxi+Wang%2C+Haonan+Li%2C+Minghao+Wu%2C+Yuxia+Wang%2C+Xudong+Han%2C+Chiyu+Zhang%2C+and+Timothy+Baldwin%2E+2024%2E+Demystifying+instruction+mixing+for+fine%2Dtuning+large+language+models%2E+In+Proceedings+of+the+62nd+Ann)
40. Weixuan Wang, Minghao Wu, Barry Haddow, and Alexandra Birch. 2025. Hbo: Hierarchical balancing optimization for fine-tuning large language models. Preprint, arXiv:2505.12300. — [arXiv:2505.12300](https://arxiv.org/abs/2505.12300) [find](https://scholar.google.com/scholar?q=Weixuan+Wang%2C+Minghao+Wu%2C+Barry+Haddow%2C+and+Alexandra+Birch%2E+2025%2E+Hbo%3A+Hierarchical+balancing+optimization+for+fine%2Dtuning+large+language+models%2E+Preprint%2C+arXiv%3A2505%2E12300%2E+%E2%80%94+%5BarXiv%3A2505%2E12300%5D%28http)
41. Xinyi Wang, Yulia Tsvetkov, and Graham Neubig. 2020. Balancing training for multilingual neural machine translation. In ACL. [find](https://scholar.google.com/scholar?q=Xinyi+Wang%2C+Yulia+Tsvetkov%2C+and+Graham+Neubig%2E+2020%2E+Balancing+training+for+multilingual+neural+machine+translation%2E+In+ACL%2E)
42. Yizhong Wang, Hamish Ivison, Pradeep Dasigi, Jack Hessel, Tushar Khot, Khyathi Chandu, David Wadden, Kelsey MacMillan, Noah A Smith, Iz Beltagy, and Hannaneh Hajishirzi. 2023. [How far can camels go? exploring the state of instruction tuning on open resources](https://md.research.elamur.ai/free/en/papers/wang-2023-far-camels-go-exploring-state/). In NeurIPS.
43. Jason Wei, Maarten Bosma, Vincent Zhao, Kelvin Guu, Adams Wei Yu, Brian Lester, Nan Du, Andrew M. Dai, and Quoc V Le. 2022. [Finetuned language models are zero-shot learners](https://md.research.elamur.ai/free/en/papers/wei-2021-finetuned-language-models-zero-shot/). In ICLR.
44. Minghao Wu, Yitong Li, Meng Zhang, Liangyou Li, Gholamreza Haffari, and Qun Liu. 2021. Uncertainty-aware balancing for multilingual and multi-domain neural machine translation training. In EMNLP. [find](https://scholar.google.com/scholar?q=Minghao+Wu%2C+Yitong+Li%2C+Meng+Zhang%2C+Liangyou+Li%2C+Gholamreza+Haffari%2C+and+Qun+Liu%2E+2021%2E+Uncertainty%2Daware+balancing+for+multilingual+and+multi%2Ddomain+neural+machine+translation+training%2E+In+EMNLP%2E)
45. Minghao Wu, Thuy-Trang Vu, Lizhen Qu, and Reza Haf. 2024. Mixture-of-skills: Learning to optimize data usage for fine-tuning large language models. In EMNLP. [find](https://scholar.google.com/scholar?q=Minghao+Wu%2C+Thuy%2DTrang+Vu%2C+Lizhen+Qu%2C+and+Reza+Haf%2E+2024%2E+Mixture%2Dof%2Dskills%3A+Learning+to+optimize+data+usage+for+fine%2Dtuning+large+language+models%2E+In+EMNLP%2E)
46. Yuxin Xiao, Shujian Zhang, Wenxuan Zhou, Marzyeh Ghassemi, and Sanqiang Zhao. 2025. [Sftmix: Elevating language model instruction tuning with mixup recipe](https://md.research.elamur.ai/free/en/papers/xiao-2026-sftmix-elevating-language-model-instruction/). Preprint, arXiv:2410.05248.
47. Sang Michael Xie, Hieu Pham, Xuanyi Dong, Nan Du, Hanxiao Liu, Yifeng Lu, Percy Liang, Quoc V. Le, Tengyu Ma, and Adams Wei Yu. 2023. Doremi: Optimizing data mixtures speeds up language model pretraining. Preprint, arXiv:2305.10429. [find](https://scholar.google.com/scholar?q=Sang+Michael+Xie%2C+Hieu+Pham%2C+Xuanyi+Dong%2C+Nan+Du%2C+Hanxiao+Liu%2C+Yifeng+Lu%2C+Percy+Liang%2C+Quoc+V%2E+Le%2C+Tengyu+Ma%2C+and+Adams+Wei+Yu%2E+2023%2E+Doremi%3A+Optimizing+data+mixtures+speeds+up+language+model+pretrain)
48. Can Xu, Qingfeng Sun, Kai Zheng, Xiubo Geng, Pu Zhao, Jiazhan Feng, Chongyang Tao, Qingwei Lin, and Daxin Jiang. 2024. [WizardLM: Empowering large pre-trained language models to follow complex instructions](https://md.research.elamur.ai/free/en/papers/xu-2023-wizardlm-empowering-large-pre-trained/). In ICLR.
49. An Yang, Anfeng Li, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chang Gao, Chengen Huang, Chenxu Lv, Chujie Zheng, Dayiheng Liu, Fan Zhou, Fei Huang, Feng Hu, Hao Ge, Haoran Wei, Huan Lin, Jialong Tang, and 41 others. 2025. [Qwen3 technical report](https://md.research.elamur.ai/free/en/papers/yang-2025-qwen3-technical-report/). Preprint, arXiv:2505.09388.
50. Wenting Zhao, Xiang Ren, Jack Hessel, Claire Cardie, Yejin Choi, and Yuntian Deng. 2024. [Wildchat: 1m chatGPT interaction logs in the wild](https://md.research.elamur.ai/free/en/papers/zhao-2024-wildchat-1m-chatgpt-interaction-logs/). In The Twelfth International Conference on Learning Representations.
51. Chunting Zhou, Pengfei Liu, Puxin Xu, Srinivasan Iyer, Jiao Sun, Yuning Mao, Xuezhe Ma, Avia Efrat, Ping Yu, LILI YU, Susan Zhang, Gargi Ghosh, Mike Lewis, Luke Zettlemoyer, and Omer Levy. 2023a. Lima: Less is more for alignment. In NeurIPS. [find](https://scholar.google.com/scholar?q=Chunting+Zhou%2C+Pengfei+Liu%2C+Puxin+Xu%2C+Srinivasan+Iyer%2C+Jiao+Sun%2C+Yuning+Mao%2C+Xuezhe+Ma%2C+Avia+Efrat%2C+Ping+Yu%2C+LILI+YU%2C+Susan+Zhang%2C+Gargi+Ghosh%2C+Mike+Lewis%2C+Luke+Zettlemoyer%2C+and+Omer+Levy%2E+2023a%2E+Lima)
52. Jeffrey Zhou, Tianjian Lu, Swaroop Mishra, Siddhartha Brahma, Sujoy Basu, Yi Luan, Denny Zhou, and Le Hou. 2023b. [Instruction-following evaluation for large language models](https://md.research.elamur.ai/free/en/papers/zhou-2023-instruction-following-evaluation-large-language/). Preprint, arXiv:2311.07911.
53. Tong Zhu, Daize Dong, Xiaoye Qu, Jiacheng Ruan, Wenliang Chen, and Yu Cheng. 2025. Dynamic data mixing maximizes instruction tuning for mixture-of-experts. In NAACL. [find](https://scholar.google.com/scholar?q=Tong+Zhu%2C+Daize+Dong%2C+Xiaoye+Qu%2C+Jiacheng+Ruan%2C+Wenliang+Chen%2C+and+Yu+Cheng%2E+2025%2E+Dynamic+data+mixing+maximizes+instruction+tuning+for+mixture%2Dof%2Dexperts%2E+In+NAACL%2E)

*Updated 1 November 2026*