FNF: Functional Network Fingerprint for Large Language Models

2026-01-30

AI Summary

This work proposes a training-free, sample-efficient functional network fingerprinting method to protect the intellectual property of open-source large language models and prevent unauthorized replication. By analyzing the consistency of neuron activation patterns across a small set of inputs, the method constructs a functional fingerprint that enables reliable attribution of suspect models. The key innovation lies in leveraging functional network activity consistency to support cross-architecture and cross-scale model comparison—an approach shown to be robust against common modifications such as fine-tuning, pruning, and parameter reordering. Experimental results demonstrate that the technique achieves high-accuracy model provenance identification with only a minimal number of query samples.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Computer Vision: Large Vision Models
• Natural Language Processing: (Large) Language Models

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

The development of large language models (LLMs) is costly and has significant commercial value. Consequently, preventing unauthorized appropriation of open-source LLMs and protecting developers'intellectual property rights have become critical challenges. In this work, we propose the Functional Network Fingerprint (FNF), a training-free, sample-efficient method for detecting whether a suspect LLM is derived from a victim model, based on the consistency between their functional network activity. We demonstrate that models that share a common origin, even with differences in scale or architecture, exhibit highly consistent patterns of neuronal activity within their functional networks across diverse input samples. In contrast, models trained independently on distinct data or with different objectives fail to preserve such activity alignment. Unlike conventional approaches, our method requires only a few samples for verification, preserves model utility, and remains robust to common model modifications (such as fine-tuning, pruning, and parameter permutation), as well as to comparisons across diverse architectures and dimensionalities. FNF thus provides model owners and third parties with a simple, non-invasive, and effective tool for protecting LLM intellectual property. The code is available at https://github.com/WhatAboutMyStar/LLM_ACTIVATION.

Problem

• intellectual property protection
• large language models
• model theft
• functional network fingerprint
• LLM security

Innovation

• Functional Network Fingerprint
• LLM Intellectual Property
• Training-free Verification
• Neuronal Activity Consistency
• Model Fingerprinting