Universal Neurons in GPT-2: Emergence, Persistence, and Functional Impact

• URL: http://arxiv.org/abs/2508.00903v1
• Date: Mon, 28 Jul 2025 22:45:33 GMT
• Title: Universal Neurons in GPT-2: Emergence, Persistence, and Functional Impact
• Abstract summary: We investigate the phenomenon of neuron universality in independently trained GPT-2 Small models.<n>We identify universal neurons through pairwise correlation analysis of activations over a dataset of 5 million tokens.<n>We quantify neuron persistence, demonstrating high stability of universal neurons across training checkpoints.
• Score: 3.6335172274433414
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: We investigate the phenomenon of neuron universality in independently trained GPT-2 Small models, examining how these universal neurons-neurons with consistently correlated activations across models-emerge and evolve throughout training. By analyzing five GPT-2 models at three checkpoints (100k, 200k, 300k steps), we identify universal neurons through pairwise correlation analysis of activations over a dataset of 5 million tokens. Ablation experiments reveal significant functional impacts of universal neurons on model predictions, measured via loss and KL divergence. Additionally, we quantify neuron persistence, demonstrating high stability of universal neurons across training checkpoints, particularly in deeper layers. These findings suggest stable and universal representational structures emerge during neural network training.

Related papers

• Topological Neural Dynamics: A Neuron-wise Framework for Sequence Modeling [55.07411490538404]
  Topological Neural Dynamics (TND) is a sequence modeling framework that shifts from layer-wise to neuron-wise dynamics.<n>TND represents a neural system as a directed neuron graph, an interaction operator, and a local dynamics function.<n>We evaluate TND as a case study on a behavior cloning task in single-player Pong.
  arXiv  Detail & Related papers  (2026-06-19T10:18:24Z)
• Neuron Populations Exhibit Divergent Selectivity with Scale [46.27365621560269]
  We study Rosetta Neurons, a class of neurons whose activation patterns are similar across independently trained models.<n>In separate analyses of language models up to 30B parameters and vision models up to 5B parameters, we observe that the population of Rosetta Neurons follows a sublinear power law in model size.<n>We further observe a Neuron Polarization Effect: Rosetta Neurons become more selective and increasingly monosemantic with scale.
  arXiv  Detail & Related papers  (2026-06-02T17:59:52Z)
• Neuronal Group Communication for Efficient Neural representation [85.36421257648294]
  This paper addresses the question of how to build large neural systems that learn efficient, modular, and interpretable representations.<n>We propose Neuronal Group Communication (NGC), a theory-driven framework that reimagines a neural network as a dynamical system of interacting neuronal groups.<n>NGC treats weights as transient interactions between embedding-like neuronal states, with neural computation unfolding through iterative communication among groups of neurons.
  arXiv  Detail & Related papers  (2025-10-19T14:23:35Z)
• Langevin Flows for Modeling Neural Latent Dynamics [81.81271685018284]
  We introduce LangevinFlow, a sequential Variational Auto-Encoder where the time evolution of latent variables is governed by the underdamped Langevin equation.<n>Our approach incorporates physical priors -- such as inertia, damping, a learned potential function, and forces -- to represent both autonomous and non-autonomous processes in neural systems.<n>Our method outperforms state-of-the-art baselines on synthetic neural populations generated by a Lorenz attractor.
  arXiv  Detail & Related papers  (2025-07-15T17:57:48Z)
• NOBLE -- Neural Operator with Biologically-informed Latent Embeddings to Capture Experimental Variability in Biological Neuron Models [68.89389652724378]
  NOBLE is a neural operator framework that learns a mapping from a continuous frequency-modulated embedding of interpretable neuron features to the somatic voltage response induced by current injection.<n>It predicts distributions of neural dynamics accounting for the intrinsic experimental variability.<n>NOBLE is the first scaled-up deep learning framework validated on real experimental data.
  arXiv  Detail & Related papers  (2025-06-05T01:01:18Z)
• Single-neuron deep generative model uncovers underlying physics of neuronal activity in Ca imaging data [0.0]
  We propose a novel framework for single-neuron representation learning using autoregressive variational autoencoders (AVAEs)<n>Our approach embeds individual neurons' signals into a reduced-dimensional space without the need for spike inference algorithms.<n>The AVAE excels over traditional linear methods by generating more informative and discriminative latent representations.
  arXiv  Detail & Related papers  (2025-01-24T16:33:52Z)
• Neuron-Level Differentiation of Memorization and Generalization in Large Language Models [9.504942958632384]
  We investigate how Large Language Models distinguish between memorization and generalization at the neuron level.<n>Experiments on both a GPT-2 model trained from scratch and a pretrained LLaMA-3.2 model fine-tuned with LoRA show consistent neuron-level specialization.
  arXiv  Detail & Related papers  (2024-12-24T15:28:56Z)
• Confidence Regulation Neurons in Language Models [91.90337752432075]
  This study investigates the mechanisms by which large language models represent and regulate uncertainty in next-token predictions. Entropy neurons are characterized by an unusually high weight norm and influence the final layer normalization (LayerNorm) scale to effectively scale down the logits. token frequency neurons, which we describe here for the first time, boost or suppress each token's logit proportionally to its log frequency, thereby shifting the output distribution towards or away from the unigram distribution.
  arXiv  Detail & Related papers  (2024-06-24T01:31:03Z)
• Universal Neurons in GPT2 Language Models [4.9892471449871305]
  We study the universality of individual neurons across GPT2 models trained from different initial random seeds. We find that 1-5% of neurons are universal, that is, pairs of neurons which consistently activate on the same inputs.
  arXiv  Detail & Related papers  (2024-01-22T18:11:01Z)
• The Neuron as a Direct Data-Driven Controller [43.8450722109081]
  This study extends the current normative models, which primarily optimize prediction, by conceptualizing neurons as optimal feedback controllers. We model neurons as biologically feasible controllers which implicitly identify loop dynamics, infer latent states and optimize control. Our model presents a significant departure from the traditional, feedforward, instant-response McCulloch-Pitts-Rosenblatt neuron, offering a novel and biologically-informed fundamental unit for constructing neural networks.
  arXiv  Detail & Related papers  (2024-01-03T01:24:10Z)
• The Expressive Leaky Memory Neuron: an Efficient and Expressive Phenomenological Neuron Model Can Solve Long-Horizon Tasks [64.08042492426992]
  We introduce the Expressive Memory (ELM) neuron model, a biologically inspired model of a cortical neuron. Our ELM neuron can accurately match the aforementioned input-output relationship with under ten thousand trainable parameters. We evaluate it on various tasks with demanding temporal structures, including the Long Range Arena (LRA) datasets.
  arXiv  Detail & Related papers  (2023-06-14T13:34:13Z)
• Generalization of generative model for neuronal ensemble inference method [0.0]
  In this study, we extend the range of the variable for expressing the neuronal state, and generalize the likelihood of the model for extended variables. This generalization without restriction of the binary input enables us to perform soft clustering and apply the method to non-stationary neuroactivity data.
  arXiv  Detail & Related papers  (2022-11-07T07:58:29Z)
• STNDT: Modeling Neural Population Activity with a Spatiotemporal Transformer [19.329190789275565]
  We introduce SpatioTemporal Neural Data Transformer (STNDT), an NDT-based architecture that explicitly models responses of individual neurons. We show that our model achieves state-of-the-art performance on ensemble level in estimating neural activities across four neural datasets.
  arXiv  Detail & Related papers  (2022-06-09T18:54:23Z)
• Continuous Learning and Adaptation with Membrane Potential and Activation Threshold Homeostasis [91.3755431537592]
  This paper presents the Membrane Potential and Activation Threshold Homeostasis (MPATH) neuron model. The model allows neurons to maintain a form of dynamic equilibrium by automatically regulating their activity when presented with input. Experiments demonstrate the model's ability to adapt to and continually learn from its input.
  arXiv  Detail & Related papers  (2021-04-22T04:01:32Z)
• The Neural Coding Framework for Learning Generative Models [91.0357317238509]
  We propose a novel neural generative model inspired by the theory of predictive processing in the brain. In a similar way, artificial neurons in our generative model predict what neighboring neurons will do, and adjust their parameters based on how well the predictions matched reality.
  arXiv  Detail & Related papers  (2020-12-07T01:20:38Z)

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.