Skip to main content
Search

Augusta University

1 related articles

Subscribe to updates
  1. Watermark audits can be erased at inference: Logit Laundering lifts detection p-values 513-fold on average while keeping 56.8% of the copyright-knowledge gain

    Data-use watermark auditing relies on statistical signals detectable in model outputs, and a non-compliant model provider can weaken those signals by rewriting the output distribution at inference time. Logit Laundering subtracts an estimated watermark-related logit bias from the base model's logits without changing model parameters, and scales the suppression by next-token prediction confidence. Across five foundation models and several mainstream auditing methods, detection p-values rise 513-fold on average and the signals become statistically insignificant, while a 56.8% average relative gain in copyrighted knowledge over the original foundation models is retained.

You’re all caught up