English

Antidistillation Fingerprinting

Machine Learning 2026-05-18 v2 Artificial Intelligence Computation and Language

Abstract

Model distillation enables efficient emulation of frontier large language models (LLMs), creating a need for robust mechanisms to detect when a third-party student model has trained on a teacher model's outputs. However, existing fingerprinting techniques that could be used to detect such distillation rely on heuristic perturbations that impose a steep trade-off between generation quality and fingerprinting strength, often requiring significant degradation of utility to ensure the fingerprint is effectively internalized by the student. We introduce antidistillation fingerprinting (ADFP), a principled approach that aligns the fingerprinting objective with the student's learning dynamics. Building upon the gradient-based framework of antidistillation sampling, ADFP utilizes a proxy model to identify and sample tokens that directly maximize the expected detectability of the fingerprint in the student after fine-tuning, rather than relying on the incidental absorption of the un-targeted biases of a more naive watermark. Experiments on GSM8K, OASST1, and MBPP demonstrate that ADFP achieves a significant Pareto improvement over state-of-the-art baselines, yielding stronger detection confidence with minimal impact on utility across mathematical reasoning, dialogue, and code generation, even when the student model's architecture is unknown.

Keywords

Cite

@article{arxiv.2602.03812,
  title  = {Antidistillation Fingerprinting},
  author = {Yixuan Even Xu and John Kirchenbauer and Yash Savani and Asher Trockman and Alexander Robey and Tom Goldstein and Fei Fang and J. Zico Kolter},
  journal= {arXiv preprint arXiv:2602.03812},
  year   = {2026}
}

Comments

28 pages, 13 figures, ICML 2026

R2 v1 2026-07-01T09:34:45.608Z