Detecting Non-Membership in LLM Training Data via Rank Correlations
Fuente:
arXiv
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908916024082432 |
|---|---|
| author | Shetty, Pranav Haque, Mirazul Ma, Zhiqiang Liu, Xiaomo |
| author_facet | Shetty, Pranav Haque, Mirazul Ma, Zhiqiang Liu, Xiaomo |
| contents | As large language models (LLMs) are trained on increasingly vast and opaque text corpora, determining which data contributed to training has become essential for copyright enforcement, compliance auditing, and user trust. While prior work focuses on detecting whether a dataset was used in training (membership inference), the complementary problem -- verifying that a dataset was not used -- has received little attention. We address this gap by introducing PRISM, a test that detects dataset-level non-membership using only grey-box access to model logits. Our key insight is that two models that have not seen a dataset exhibit higher rank correlation in their normalized token log probabilities than when one model has been trained on that data. Using this observation, we construct a correlation-based test that detects non-membership. Empirically, PRISM reliably rules out membership in training data across all datasets tested while avoiding false positives, thus offering a framework for verifying that specific datasets were excluded from LLM training. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_22707 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Detecting Non-Membership in LLM Training Data via Rank Correlations Shetty, Pranav Haque, Mirazul Ma, Zhiqiang Liu, Xiaomo Computation and Language As large language models (LLMs) are trained on increasingly vast and opaque text corpora, determining which data contributed to training has become essential for copyright enforcement, compliance auditing, and user trust. While prior work focuses on detecting whether a dataset was used in training (membership inference), the complementary problem -- verifying that a dataset was not used -- has received little attention. We address this gap by introducing PRISM, a test that detects dataset-level non-membership using only grey-box access to model logits. Our key insight is that two models that have not seen a dataset exhibit higher rank correlation in their normalized token log probabilities than when one model has been trained on that data. Using this observation, we construct a correlation-based test that detects non-membership. Empirically, PRISM reliably rules out membership in training data across all datasets tested while avoiding false positives, thus offering a framework for verifying that specific datasets were excluded from LLM training. |
| title | Detecting Non-Membership in LLM Training Data via Rank Correlations |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2603.22707 |