# Transferable text data distillation by trajectory matching

arXiv: https://arxiv.org/abs/2504.09818 · HTML: https://mlsanity.com/paper/2504.09818/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 24 of its 31 references that are in the map.
- First posted on arXiv 14 Apr 2025.

## Builds on (most cited first, 24 of 24)

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Chain-of-Thought Prompting Elicits Reasoning in Large Language Models](https://mlsanity.com/paper/2201.11903/) (2022)
- [Measuring Massive Multitask Language Understanding](https://mlsanity.com/paper/2009.03300/) (2020)
- [Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge](https://mlsanity.com/paper/1803.05457/) (2018)
- [OPT: Open Pre-trained Transformer Language Models](https://mlsanity.com/paper/2205.01068/) (2022)
- [PaLM 2 Technical Report](https://mlsanity.com/paper/2305.10403/) (2023)
- [The Flan Collection: Designing Data and Methods for Effective Instruction Tuning](https://mlsanity.com/paper/2301.13688/) (2023)
- [OpenAssistant Conversations -- Democratizing Large Language Model Alignment](https://mlsanity.com/paper/2304.07327/) (2023)
- [Dataset Condensation with Gradient Matching](https://mlsanity.com/paper/2006.05929/) (2020)
- [Dataset Distillation by Matching Training Trajectories](https://mlsanity.com/paper/2203.11932/) (2022)
- [Dataset Condensation with Distribution Matching](https://mlsanity.com/paper/2110.04181/) (2021)
- [Data Selection for Language Models via Importance Resampling](https://mlsanity.com/paper/2302.03169/) (2023)
- [Dataset Condensation with Differentiable Siamese Augmentation](https://mlsanity.com/paper/2102.08259/) (2021)
- [Soft-Label Dataset Distillation and Text Dataset Distillation](https://mlsanity.com/paper/1910.02551/) (2019)
- [Flexible Dataset Distillation: Learn Labels Instead of Images](https://mlsanity.com/paper/2006.08572/) (2020)
- [Data Distillation: A Survey](https://mlsanity.com/paper/2301.04272/) (2023)
- [Evaluating Instruction-Tuned Large Language Models on Code Comprehension and Generation](https://mlsanity.com/paper/2308.01240/) (2023)
- [Dataset Quantization](https://mlsanity.com/paper/2308.10524/) (2023)
- [Data Distillation for Text Classification](https://mlsanity.com/paper/2104.08448/) (2021)
- [DiLM: Distilling Dataset into Language Model for Text-level Dataset Distillation](https://mlsanity.com/paper/2404.00264/) (2024)
- [The Trade-offs of Domain Adaptation for Neural Language Models](https://mlsanity.com/paper/2109.10274/) (2021)
- [On the Complementarity of Data Selection and Fine Tuning for Domain Adaptation](https://mlsanity.com/paper/2109.07591/) (2021)

## Built on by (most cited first, 1 of 1)

- [Is Human-Readable Text Necessary for Effective LLM Fine-Tuning?](https://mlsanity.com/paper/2609.35868/) (2026)

## Abstract (arXiv)

In the realm of large language model (LLM), as the size of large models increases, it also brings higher training costs. There is a urgent need to minimize the data size in LLM training. Compared with data selection method, the data distillation method aims to synthesize a small number of data samples to achieve the training effect of the full data set and has better flexibility. Despite its successes in computer vision, the discreteness of text data has hitherto stymied its exploration in natural language processing (NLP). In this work, we proposed a method that involves learning pseudo prompt data based on trajectory matching and finding its nearest neighbor ID to achieve cross-architecture transfer. During the distillation process, we introduce a regularization loss to improve the robustness of our distilled data. To our best knowledge, this is the first data distillation work suitable for text generation tasks such as instruction tuning. Evaluations on two benchmarks, including ARC-Easy and MMLU instruction tuning datasets, established the superiority of our distillation approach over the SOTA data selection method LESS. Furthermore, our method demonstrates a good transferability over LLM structures (i.e., OPT to Llama).