Tech Meridian ← LIVE FEED
PROMY MERIDIAN RU

RESEARCH · RESEARCH · #3185

Embedding-space fingerprinting recovers fine-tune domain from LoRA artifacts

This paper (arXiv:2611.07610v1) studies whether the domain of a LoRA fine-tune can be recovered from the model artifact alone. Using Qwen3.5-0.8B as the base, the authors produce 36 LoRA fine-tunes across 18 public datasets in five domains and extract three feature families (weight deltas, internal activations, and outputs) under a pre-registered leave-one-dataset-out (LODO) protocol; a classifier identifies the domain well above chance from gray-box activations (61%) and black-box outputs (64%), while white-box weight statistics reach 42%, and the black-box result holds even when only generated text is used.

KEY POINTS

  1. This paper (arXiv:2611.07610v1) studies whether the domain of a LoRA fine-tune can be recovered from the model artifact alone.
  2. Using Qwen3.5-0.8B as the base, the authors produce 36 LoRA fine-tunes across 18 public datasets in five domains and extract three feature families (weight deltas, internal activations, and outputs) under a pre-registered leave-one-dataset-out (LODO) protocol; a classifier identifies the domain well above chance from gray-box activations (61%) and black-box outputs (64%), while white-box weight statistics reach 42%, and the black-box result holds even when only generated text is used.
  3. Shows that fine-tuned LoRA artifacts can leak the training-data domain from weights, activations, or outputs, with implications for privacy, model sharing, and defenses such as watermarking.

WHY IT MATTERS

Shows that fine-tuned LoRA artifacts can leak the training-data domain from weights, activations, or outputs, with implications for privacy, model sharing, and defenses such as watermarking.

SOURCES & TIMELINE

1