From Unfamiliar to Familiar: Detecting Pre-training Data via Gradient Deviations in Large Language Models

• URL: http://arxiv.org/abs/2603.04828v1
• Date: Thu, 05 Mar 2026 05:21:51 GMT
• Title: From Unfamiliar to Familiar: Detecting Pre-training Data via Gradient Deviations in Large Language Models
• Authors: Ruiqi Zhang, Lingxiang Wang, Hainan Zhang, Zhiming Zheng, Yanyan Lan,
• Abstract summary: Existing methods mainly focus on the likelihood-based statistical features or signals before and after fine-tuning.<n>We propose GDS, a method that identifies pre-training data by probing Gradient Deviation Scores of target samples.<n>GDS achieves state-of-the-art performance with significantly improved cross-dataset transferability over strong baselines.
• Score: 24.38939297043815
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Pre-training data detection for LLMs is essential for addressing copyright concerns and mitigating benchmark contamination. Existing methods mainly focus on the likelihood-based statistical features or heuristic signals before and after fine-tuning, but the former are susceptible to word frequency bias in corpora, and the latter strongly depend on the similarity of fine-tuning data. From an optimization perspective, we observe that during training, samples transition from unfamiliar to familiar in a manner reflected by systematic differences in gradient behavior. Familiar samples exhibit smaller update magnitudes, distinct update locations in model components, and more sharply activated neurons. Based on this insight, we propose GDS, a method that identifies pre-training data by probing Gradient Deviation Scores of target samples. Specifically, we first represent each sample using gradient profiles that capture the magnitude, location, and concentration of parameter updates across FFN and Attention modules, revealing consistent distinctions between member and non-member data. These features are then fed into a lightweight classifier to perform binary membership inference. Experiments on five public datasets show that GDS achieves state-of-the-art performance with significantly improved cross-dataset transferability over strong baselines. Further interpretability analyse show gradient feature distribution differences, enabling practical and scalable pre-training data detection.

Related papers

• Exploiting Inter-Sample Information for Long-tailed Out-of-Distribution Detection [7.0229899259286945]
  We show that exploiting inter-sample relationships can significantly improve OOD detection in long-tailed recognition of vision datasets.<n>Our method outperforms the state-of-the-art approaches by a large margin in terms of FPR and tail-class ID classification accuracy.
  arXiv  Detail & Related papers  (2025-11-20T03:31:37Z)
• Fine-tuning can Help Detect Pretraining Data from Large Language Models [7.7209640786782385]
  Current methods differentiate members and non-members by designing scoring functions, like Perplexity and Min-k%.<n>We introduce a novel and effective method termed Fine-tuned Score Deviation(FSD), which improves the performance of current scoring functions for pretraining data detection.<n>In particular, we propose to measure the deviation distance of current scores after fine-tuning on a small amount of unseen data within the same domain.
  arXiv  Detail & Related papers  (2024-10-09T15:36:42Z)
• Downstream-Pretext Domain Knowledge Traceback for Active Learning [138.02530777915362]
  We propose a downstream-pretext domain knowledge traceback (DOKT) method that traces the data interactions of downstream knowledge and pre-training guidance. DOKT consists of a traceback diversity indicator and a domain-based uncertainty estimator. Experiments conducted on ten datasets show that our model outperforms other state-of-the-art methods.
  arXiv  Detail & Related papers  (2024-07-20T01:34:13Z)
• Impact of Noisy Supervision in Foundation Model Learning [91.56591923244943]
  This paper is the first work to comprehensively understand and analyze the nature of noise in pre-training datasets.<n>We propose a tuning method (NMTune) to affine the feature space to mitigate the malignant effect of noise and improve generalization.
  arXiv  Detail & Related papers  (2024-03-11T16:22:41Z)
• Assessing Neural Network Representations During Training Using Noise-Resilient Diffusion Spectral Entropy [55.014926694758195]
  Entropy and mutual information in neural networks provide rich information on the learning process. We leverage data geometry to access the underlying manifold and reliably compute these information-theoretic measures. We show that they form noise-resistant measures of intrinsic dimensionality and relationship strength in high-dimensional simulated data.
  arXiv  Detail & Related papers  (2023-12-04T01:32:42Z)
• Cluster-level pseudo-labelling for source-free cross-domain facial expression recognition [94.56304526014875]
  We propose the first Source-Free Unsupervised Domain Adaptation (SFUDA) method for Facial Expression Recognition (FER) Our method exploits self-supervised pretraining to learn good feature representations from the target data. We validate the effectiveness of our method in four adaptation setups, proving that it consistently outperforms existing SFUDA methods when applied to FER.
  arXiv  Detail & Related papers  (2022-10-11T08:24:50Z)
• Imputation-Free Learning from Incomplete Observations [73.15386629370111]
  We introduce the importance of guided gradient descent (IGSGD) method to train inference from inputs containing missing values without imputation. We employ reinforcement learning (RL) to adjust the gradients used to train the models via back-propagation. Our imputation-free predictions outperform the traditional two-step imputation-based predictions using state-of-the-art imputation methods.
  arXiv  Detail & Related papers  (2021-07-05T12:44:39Z)
• Omni-supervised Facial Expression Recognition via Distilled Data [120.11782405714234]
  We propose omni-supervised learning to exploit reliable samples in a large amount of unlabeled data for network training. We experimentally verify that the new dataset can significantly improve the ability of the learned FER model. To tackle this, we propose to apply a dataset distillation strategy to compress the created dataset into several informative class-wise images.
  arXiv  Detail & Related papers  (2020-05-18T09:36:51Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.