> The watermark doesn't change the meaning, quality, or readability of the output
how?
Here's the paper describing the technique: https://www.nature.com/articles/s41586-024-08025-4
It manipulates the PRNG seed in a systematic way, keeping the same token sampling distribution.
https://www.nature.com/articles/s41586-024-08025-4
Here's the paper describing the technique: https://www.nature.com/articles/s41586-024-08025-4