#
pplm
Here are 2 public repositories matching this topic...
A research implementation of statistical text watermarking for large language models using Plug and Play Language Models (PPLM). This system enables detectable watermark embedding through direct logit perturbation during inference without modifying the base model weights.
natural-language-processing transformers pytorch ai-safety statistical-detection gpt-2 pplm llm text-watermarking ai-content-detection
-
Updated
Feb 17, 2026 - Python
Add this topic to your repo
To associate your repository with the pplm topic, visit your repo's landing page and select "manage topics."