On the Reliability of Watermarks for Large Language Models
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2023
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866916231167082496 |
|---|---|
| author | Kirchenbauer, John Geiping, Jonas Wen, Yuxin Shu, Manli Saifullah, Khalid Kong, Kezhi Fernando, Kasun Saha, Aniruddha Goldblum, Micah Goldstein, Tom |
| author_facet | Kirchenbauer, John Geiping, Jonas Wen, Yuxin Shu, Manli Saifullah, Khalid Kong, Kezhi Fernando, Kasun Saha, Aniruddha Goldblum, Micah Goldstein, Tom |
| contents | As LLMs become commonplace, machine-generated text has the potential to flood the internet with spam, social media bots, and valueless content. Watermarking is a simple and effective strategy for mitigating such harms by enabling the detection and documentation of LLM-generated text. Yet a crucial question remains: How reliable is watermarking in realistic settings in the wild? There, watermarked text may be modified to suit a user's needs, or entirely rewritten to avoid detection. We study the robustness of watermarked text after it is re-written by humans, paraphrased by a non-watermarked LLM, or mixed into a longer hand-written document. We find that watermarks remain detectable even after human and machine paraphrasing. While these attacks dilute the strength of the watermark, paraphrases are statistically likely to leak n-grams or even longer fragments of the original text, resulting in high-confidence detections when enough tokens are observed. For example, after strong human paraphrasing the watermark is detectable after observing 800 tokens on average, when setting a 1e-5 false positive rate. We also consider a range of new detection schemes that are sensitive to short spans of watermarked text embedded inside a large document, and we compare the robustness of watermarking to other kinds of detectors. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2306_04634 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | On the Reliability of Watermarks for Large Language Models Kirchenbauer, John Geiping, Jonas Wen, Yuxin Shu, Manli Saifullah, Khalid Kong, Kezhi Fernando, Kasun Saha, Aniruddha Goldblum, Micah Goldstein, Tom Machine Learning Computation and Language Cryptography and Security As LLMs become commonplace, machine-generated text has the potential to flood the internet with spam, social media bots, and valueless content. Watermarking is a simple and effective strategy for mitigating such harms by enabling the detection and documentation of LLM-generated text. Yet a crucial question remains: How reliable is watermarking in realistic settings in the wild? There, watermarked text may be modified to suit a user's needs, or entirely rewritten to avoid detection. We study the robustness of watermarked text after it is re-written by humans, paraphrased by a non-watermarked LLM, or mixed into a longer hand-written document. We find that watermarks remain detectable even after human and machine paraphrasing. While these attacks dilute the strength of the watermark, paraphrases are statistically likely to leak n-grams or even longer fragments of the original text, resulting in high-confidence detections when enough tokens are observed. For example, after strong human paraphrasing the watermark is detectable after observing 800 tokens on average, when setting a 1e-5 false positive rate. We also consider a range of new detection schemes that are sensitive to short spans of watermarked text embedded inside a large document, and we compare the robustness of watermarking to other kinds of detectors. |
| title | On the Reliability of Watermarks for Large Language Models |
| topic | Machine Learning Computation and Language Cryptography and Security |
| url | https://arxiv.org/abs/2306.04634 |