Diffusion Models Preferentially Memorize Prototypical Examples or: Why Does My Diffusion Model Love Slop?

• URL: http://arxiv.org/abs/2605.30642v1
• Date: Thu, 28 May 2026 22:52:56 GMT
• Title: Diffusion Models Preferentially Memorize Prototypical Examples or: Why Does My Diffusion Model Love Slop?
• Abstract summary: We train diffusion models on strings generated according to the Random Hierarchy Model (RHM)<n>We find that samples composed of commons are preferentially memorized.<n>We identify an intermediate regime of partial memorization in which commons are learned first and subsequently overproduced during generation.
• Score: 1.3416169841532526
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Generative models have a persistent limitation: their tendency to memorize training data can create legal liabilities and erode creative diversity. Understanding which samples are memorized in whole or in part, and under what conditions, therefore remains an important open problem. Here we answer the question "Are atypical or rare samples memorized first?" in the negative. We train diffusion models on strings generated according to the production rules of the Random Hierarchy Model (RHM), and find that samples composed of common substrings are preferentially memorized. This holds true even if the training data consists of entirely unique samples, indicating that deduplication at the data point level does not provide a meaningful privacy guarantee. Correspondingly we predict, then observe, delayed memorization for fat-tailed datasets (i.e., those with more atypical samples). This effect is amplified when fat-tails are introduced into high-level production rules. These together suggest that dataset diversity, particularly at higher levels of abstraction, plays an important role in staving off memorization. Finally, we identify an intermediate regime of partial memorization in which common substrings are learned first and subsequently overproduced during generation. If training is stopped in this regime, models will exhibit the reversion-to-the-mean blandness often derided as "slop".

Related papers

• Local Coverage Governs Memorization in Diffusion Models [10.085837353469737]
  We show that memorization is governed by emphlocal data coverage.<n>We show that classes with higher intra-class sparsity (and thus lower local coverage) are more strongly memorized.
  arXiv  Detail & Related papers  (2026-06-12T12:25:13Z)
• Selective Underfitting in Diffusion Models [62.51850508420274]
  We argue that diffusion models underfit the empirical score due to training-time inductive biases.<n>Instead of underfitting the score everywhere, better diffusion models more accurately approximate the score in certain regions of input space, while underfitting it in others.<n>Our results establish that selective underfitting is essential for understanding diffusion models, yielding new, testable insights into their generalization and generative performance.
  arXiv  Detail & Related papers  (2025-10-01T19:06:20Z)
• A Closer Look on Memorization in Tabular Diffusion Model: A Data-Centric Perspective [15.33961902853653]
  We quantify memorization for each real sample based on how many generated samples are flagged as replicas.<n>Our empirical analysis reveals a heavy-tailed distribution of memorization counts.<n>We propose DynamicCut, a two-stage, model-agnostic mitigation method.
  arXiv  Detail & Related papers  (2025-05-28T13:06:00Z)
• Redistribute Ensemble Training for Mitigating Memorization in Diffusion Models [31.92526915009259]
  Diffusion models are known for their tremendous ability to generate high-quality samples.<n>Recent methods for memory mitigation have primarily addressed the issue within the context of the text modality.<n>We propose a novel method for diffusion models from the perspective of visual modality, which is more generic and fundamental for mitigating memorization.
  arXiv  Detail & Related papers  (2025-02-13T15:56:44Z)
• Demystifying Verbatim Memorization in Large Language Models [67.49068128909349]
  Large Language Models (LLMs) frequently memorize long sequences verbatim, often with serious legal and privacy implications. We develop a framework to study verbatim memorization in a controlled setting by continuing pre-training from Pythia checkpoints with injected sequences. We find that (1) non-trivial amounts of repetition are necessary for verbatim memorization to happen; (2) later (and presumably better) checkpoints are more likely to memorize verbatim sequences, even for out-of-distribution sequences.
  arXiv  Detail & Related papers  (2024-07-25T07:10:31Z)
• Understanding and Mitigating Copying in Diffusion Models [53.03978584040557]
  Images generated by diffusion models like Stable Diffusion are increasingly widespread. Recent works and even lawsuits have shown that these models are prone to replicating their training data, unbeknownst to the user.
  arXiv  Detail & Related papers  (2023-05-31T17:58:02Z)
• Emergent and Predictable Memorization in Large Language Models [23.567027014457775]
  Memorization, or the tendency of large language models to output entire sequences from their training data verbatim, is a key concern for safely deploying language models. We seek to predict which sequences will be memorized before a large model's full train-time by extrapolating the memorization behavior of lower-compute trial runs. We provide further novel discoveries on the distribution of memorization scores across models and data.
  arXiv  Detail & Related papers  (2023-04-21T17:58:31Z)
• Reducing Training Sample Memorization in GANs by Training with Memorization Rejection [80.0916819303573]
  We propose rejection memorization, a training scheme that rejects generated samples that are near-duplicates of training samples during training. Our scheme is simple, generic and can be directly applied to any GAN architecture.
  arXiv  Detail & Related papers  (2022-10-21T20:17:50Z)
• DiscrimLoss: A Universal Loss for Hard Samples and Incorrect Samples Discrimination [28.599571524763785]
  Given data with label noise (i.e., incorrect data), deep neural networks would gradually memorize the label noise and impair model performance. To relieve this issue, curriculum learning is proposed to improve model performance and generalization by ordering training samples in a meaningful sequence.
  arXiv  Detail & Related papers  (2022-08-21T13:38:55Z)
• Quantifying Memorization Across Neural Language Models [61.58529162310382]
  Large language models (LMs) have been shown to memorize parts of their training data, and when prompted appropriately, they will emit the memorized data verbatim. This is undesirable because memorization violates privacy (exposing user data), degrades utility (repeated easy-to-memorize text is often low quality), and hurts fairness (some texts are memorized over others). We describe three log-linear relationships that quantify the degree to which LMs emit memorized training data.
  arXiv  Detail & Related papers  (2022-02-15T18:48:31Z)
• Automatic Recall Machines: Internal Replay, Continual Learning and the Brain [104.38824285741248]
  Replay in neural networks involves training on sequential data with memorized samples, which counteracts forgetting of previous behavior caused by non-stationarity. We present a method where these auxiliary samples are generated on the fly, given only the model that is being trained for the assessed objective. Instead the implicit memory of learned samples within the assessed model itself is exploited.
  arXiv  Detail & Related papers  (2020-06-22T15:07:06Z)

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.