A backdoor watermark for LLMs using passthrough layers trained to output high-entropy text on a private key, with near-perfect extraction in benign settings but with layer-removal robustness contradicted by the paper's own results.
R.; Angeli, G.; Potts, C.; and Manning, C
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2024 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Task-Agnostic Language Model Watermarking via High Entropy Passthrough Layers
A backdoor watermark for LLMs using passthrough layers trained to output high-entropy text on a private key, with near-perfect extraction in benign settings but with layer-removal robustness contradicted by the paper's own results.