--- title: "DS2-Instruct — paper by Xu et al. (2026) — summary and ideas" description: "This matters because it offers a cheap, scalable way to adapt LLMs to niche domains without collecting… 1 reference, cited by 2 papers here." authors: ["Ruiyao Xu", "Noelle I. Samia", "Han Liu"] year: 2026 venue: "Findings of the Association for Computational Linguistics: EACL 2026" source: "https://aclanthology.org/2026.findings-eacl.176/" doi: "10.18653/v1/2026.findings-eacl.176" 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/xu-2026-ds2-instruct-domain-specific-data/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/xu-2026-ds2-instruct-domain-specific-data/) [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/7/) / DS2-Instruct: Domain-Specific Data Synthesis for Large Language Models Instruction Tuning # DS2-Instruct: Domain-Specific Data Synthesis for Large Language Models Instruction Tuning **Authors:** [Ruiyao Xu](https://md.research.elamur.ai/free/en/authors/ruiyao-xu/), Noelle I. Samia and [Han Liu](https://md.research.elamur.ai/free/en/authors/han-liu/) · **Year:** 2026 · **Venue:** [Findings of the Association for Computational Linguistics: EACL 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-eacl.176/) · [PDF](https://aclanthology.org/2026.findings-eacl.176.pdf) · [DOI 10.18653/v1/2026.findings-eacl.176](https://doi.org/10.18653/v1/2026.findings-eacl.176) · [OpenAlex](https://openalex.org/W7140140937) **Citations:** 0 citations (OpenAlex, 2026-10-14) · [2](https://md.research.elamur.ai/free/en/papers/xu-2026-ds2-instruct-domain-specific-data/cited-by/) in our library, of the 149,439 papers whose bibliographies we read ## Summary Instruction-tuning an LLM for a specialized field usually requires expensive expert annotation. The alternatives are domain corpora, which are often proprietary or missing, or human-written seed examples. DS2-Instruct removes these requirements. It is a zero-shot pipeline that builds domain instruction data from only a task definition. The pipeline has three stages: 1. **Keyword pool.** It grows a pool of domain keywords by repeatedly asking an LLM for both more basic prerequisite concepts and more advanced ones. It then adds terms mined from passages retrieved out of a mixed document collection. 2. **Instruction generation.** It combines these keywords with the cognitive levels of Bloom's Taxonomy to vary question types. 3. **Filtering.** It keeps only instruction–response pairs whose sampled answers agree often enough (self-consistency voting). Using Qwen2.5-72B-Instruct as the generator, the authors built datasets for seven domains, including mathematics, finance, biomedicine and logical reasoning. Smaller open models fine-tuned on this data beat both their base versions and models trained with prior synthesis methods. This matters because it offers a cheap, scalable way to adapt LLMs to niche domains without collecting any domain data first. ## Key points - **Fine-tuned models:** Llama3-8B, Mistral-7B and Qwen2.5-7B, evaluated on seven domain benchmarks such as GSM8K, Math, CFA, MedQA, PubMedQA, LogiQA and GPQA. - **Retrieval setup:** The corpus is sampled from five subsets of the Pile (ArXiv, FreeLaw, StackExchange, Wikipedia, Github). It is queried with BM25-style scoring using the task description plus random keyword subsets. - **Instruction diversity:** Instructions span six Bloom levels: Remember, Understand, Apply, Analyze, Evaluate and Create. - **Scaling:** Mistral-7B accuracy keeps rising from 1,000 to 6,000 training samples without saturating. On Qwen models of 3B, 7B and 14B parameters, the method beats zero-shot at every size, but the gains shrink as models grow. - **Ablations (Mistral-7B, four datasets):** Removing keyword expansion or retrieval, the Bloom levels, or self-consistency filtering each lowers accuracy. In one example, accuracy falls from 42.75% to 35.64% without the filtering step. *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 - [Northwestern University](https://md.research.elamur.ai/free/en/universities/northwestern-university/) — University, Evanston, United States · 41 works · 2,194 citations 📚 Shelves (4): [🖥️ 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 🧭 Topics (1): [🧠 Language models & pretraining](https://md.research.elamur.ai/free/en/topics/language-models/) · 3,320 works ## Cite Cite as: Ruiyao Xu, Noelle I. Samia and Han Liu. 2026. DS2-Instruct: Domain-Specific Data Synthesis for Large Language Models Instruction Tuning. Findings of the Association for Computational Linguistics: EACL 2026. The paper itself, not this page. BibTeX from the library's record of the work. ```bibtex @inproceedings{xu2026ds2, title = {{DS2-Instruct: Domain-Specific Data Synthesis for Large Language Models Instruction Tuning}}, author = {Ruiyao Xu, Noelle I. Samia and Han Liu}, booktitle = {Findings of the Association for Computational Linguistics: EACL 2026}, year = {2026}, doi = {10.18653/v1/2026.findings-eacl.176}, url = {https://doi.org/10.18653/v1/2026.findings-eacl.176} } ``` - [← previous DocLens: A Tool-Augmented Multi-Agent Framework for Long Visual Document Understanding](https://md.research.elamur.ai/free/en/papers/zhu-2026-doclens-tool-augmented-multi-agent/) - [next → DuFFin: A Dual-Level Fingerprinting Framework for LLMs IP Protection](https://md.research.elamur.ai/free/en/papers/yan-2026-duffin-dual-level-fingerprinting-framework/) - [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/7/) ## References 1 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet. 1. Stella Biderman, Kieran Bicheno, and Leo Gao. 2022. Datasheet for the pile. arXiv preprint arXiv:2201.07311. — [arXiv:2201.07311](https://arxiv.org/abs/2201.07311) [find](https://scholar.google.com/scholar?q=Stella+Biderman%2C+Kieran+Bicheno%2C+and+Leo+Gao%2E+2022%2E+Datasheet+for+the+pile%2E+arXiv+preprint+arXiv%3A2201%2E07311%2E+%E2%80%94+%5BarXiv%3A2201%2E07311%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F2201%2E07311%29) *Updated 31 October 2026*