Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2603.13790 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866918388029194240 |
|---|---|
| author | Dong, Victor Ye Lee, Kuan-Yun Shuai, Jiamei Liu, Shengfei Liu, Yi Jiao, Jian |
| author_facet | Dong, Victor Ye Lee, Kuan-Yun Shuai, Jiamei Liu, Shengfei Liu, Yi Jiao, Jian |
| contents | We present \emph{Greedy Information Projection} (\textsc{GIP}), a principled framework for choosing training examples for large language model fine-tuning. \textsc{GIP} casts selection as maximizing mutual information between a subset of examples and task-specific query signals, which may originate from LLM quality judgments, metadata, or other sources. The framework involves optimizing a closed-form mutual information objective defined using both data and query embeddings, naturally balancing {\it quality} and {\it diversity}. Optimizing this score is equivalent to maximizing the projection of the query embedding matrix onto the span of the selected data, which provides a geometric explanation for the co-emergence of quality and diversity. Building on this view, we employ a fast greedy matching-pursuit procedure with efficient projection-based updates. On instruction-following and mathematical reasoning datasets, \textsc{GIP} selects small subsets that match full-data fine-tuning while using only a fraction of examples and compute, unifying quality-aware and diversity-aware selection for efficient fine-tuning. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_13790 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Greedy Information Projection for LLM Data Selection Dong, Victor Ye Lee, Kuan-Yun Shuai, Jiamei Liu, Shengfei Liu, Yi Jiao, Jian Machine Learning Computation and Language We present \emph{Greedy Information Projection} (\textsc{GIP}), a principled framework for choosing training examples for large language model fine-tuning. \textsc{GIP} casts selection as maximizing mutual information between a subset of examples and task-specific query signals, which may originate from LLM quality judgments, metadata, or other sources. The framework involves optimizing a closed-form mutual information objective defined using both data and query embeddings, naturally balancing {\it quality} and {\it diversity}. Optimizing this score is equivalent to maximizing the projection of the query embedding matrix onto the span of the selected data, which provides a geometric explanation for the co-emergence of quality and diversity. Building on this view, we employ a fast greedy matching-pursuit procedure with efficient projection-based updates. On instruction-following and mathematical reasoning datasets, \textsc{GIP} selects small subsets that match full-data fine-tuning while using only a fraction of examples and compute, unifying quality-aware and diversity-aware selection for efficient fine-tuning. |
| title | Greedy Information Projection for LLM Data Selection |
| topic | Machine Learning Computation and Language |
| url | https://arxiv.org/abs/2603.13790 |