A Watermark for Large Language Models
Kirchenbauer et al., 2023, A Watermark for Large Language Models (Proceedings of the 40th International Conference on Machine Learning (ICML 2023), PMLR 202)
Subscriber deck — free during early access
1 cards in “4.2 Impact on quality of generated text” · whole deck
4.2 Impact on quality of generated text
- When does the soft watermark leave perplexity unchanged?At both extremes: if the model’s distribution is uniform, a random green list still gives uniform sampling; if all mass is on one token, the bias doesn’t change the choice. Perplexity is affected only for tokens of moderate entropy.
- Kirchenbauer et al., 2023, A Watermark for Large Language Models (Proceedings of the 40th International Conference on Machine Learning (ICML 2023), PMLR 202), §4.2
Source
- John Kirchenbauer, Jonas Geiping, Yuxin Wen, Jonathan Katz, Ian Miers and Tom Goldstein. 2023. A Watermark for Large Language Models. In Proceedings of the 40th International Conference on Machine Learning (ICML 2023), PMLR 202, pp. 17061–17084. https://proceedings.mlr.press/v202/kirchenbauer23a.html