signature-filtering

Tag

Cards List
#signature-filtering

Signature filtering: a lightweight enhancement for statistical watermark detection in large language models

arXiv cs.LG · 2026-06-18 Cached

Signature filtering is a detection-time module that improves statistical watermark detection in LLMs by learning and removing 'signature' tokens that make watermark tests unreliable, achieving large gains in detection rates while keeping false positives low.

0 favorites 0 likes
← Back to home

Submit Feedback