DEW creates a robust watermark for LLM text by applying vector-space operations to dual embeddings and hiding the signal via key-seeded random projections, showing improved detection after paraphrasing and translation.
Proceedings of the 8th Workshop on Multimedia and Security , location =
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Robust Text Watermarking for Large Language Models via Dual Semantic Embeddings
DEW creates a robust watermark for LLM text by applying vector-space operations to dual embeddings and hiding the signal via key-seeded random projections, showing improved detection after paraphrasing and translation.