Neural Network Memorization

Ask about this topic

Overview

Neural network memorization concerns how models retain specific examples or facts, from associative recall in Hopfield networks to verbatim or factual recall in language models. Researchers study capacity and training dynamics—including noisy-label fitting and grokking—and use probes, canaries, spectral and geometric measures, and controlled benchmarks to distinguish memorized information from generalizable structure. Methods such as unlearning, parameter editing, and data-reuse strategies aim to remove or manage stored information, while preserving useful performance. A central challenge is that observable recall does not always prove a training example was memorized, and reducing memorization can either protect privacy or impair useful learning.

Latest papers

CardsList

65 papers6 papers in the last four weeks, level with the four weeks before. 0.0% of all new papers.

  1. Redundancy Maximization as a Principle of Associative Memory Learning in Hopfield Networks

    Nov 4, 2025Mark Blümel, Andreas C. Schneider, Valentin Neuhaus +5Neural Network MemorizationHopfield Networks

    Complex Systems Theory, Max Planck Institute for Dynamics and Self-Organization, Göttingen, Germany · Faculty of Physics, Institute for the Dynamics of Complex Systems, University of Göttingen · Göttingen Campus Institute for Dynamics of Biological Networks, University of Göttingen, Göttingen, Germany +1

  2. Memorizing Long-tail Data Can Help Generalization Through Composition

    Oct 18, 2025Mo Zhou, Haoyang Ma, Rong GeNeural Network GeneralizationNeural Network Memorization

    University of Washington · Duke University

  3. Dynamical stability for dense patterns in attractor neural networks

    Jul 14, 2025Uri Cohen, Máté LengyelNeural Network MemorizationAssociative Memory

    Computational and Biological Learning Lab, Department of Engineering, University of Cambridge, Cambridge, UK · Center for Cognitive Computation, Department of Cognitive Science, Central European University, Budapest, Hungary

  4. Continual Memorization of Factoids in Language Models

    Nov 11, 2024Howard Chen, Jiayi Geng, Adithya Bhaskar +2Continual Learning for LLMsMemorization in Language Models

    Princeton Language and Intelligence (PLI), Princeton University

  5. Mitigating Memorization In Language Models

    Oct 3, 2024Mansi Sakarvadia, Aswathy Ajith, Arham Khan +6Memorization in Language ModelsLLM Fine-Tuning

    University of Chicago · Argonne National Laboratory · Lawrence Berkeley National Laboratory +3