Neuron-Level Differentiation of Memorization and Generalization in Large Language Models

TL;DR

This paper uncovers neuron-level distinctions between memorization and generalization in large language models, demonstrating that specific neurons are responsible for each behavior and can be manipulated at inference time to steer model responses.

Contribution

It identifies neuron subsets responsible for memorization and generalization, showing their modularity and enabling behavior control through inference-time interventions.

Findings

Neuron subsets are responsible for memorization and generalization.

Inference-time interventions can steer model behavior.

Neuron-behavior associations are consistent across models and tasks.

Abstract

We investigate how Large Language Models (LLMs) distinguish between memorization and generalization at the neuron level. Through carefully designed tasks, we identify distinct neuron subsets responsible for each behavior. 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. We further demonstrate that inference-time interventions on these neurons can steer the model's behavior toward memorization or generalization. To assess robustness, we evaluate intra-task and inter-task consistency, confirming that these neuron-behavior associations reflect generalizable patterns rather than dataset-specific artifacts. Our findings reveal modular structure in LLMs and enable controlling memorization and generalization behaviors at inference time.