We lifted 6 functions out of this paper's own repositories and ran 6 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.
| Repository | Role | Ran |
|---|---|---|
| facebookresearch/three_bricks | canonical | 4 of 4 |
| copy not recorded | — | 2 of 2 |
| Function | Status | Where it lives |
|---|---|---|
| format_prompts | Ran | facebookresearch/three_bricks/main_watermark.py pointer only (licence: NOASSERTION) · get_code("694c2ab8d8cd558c") |
| load_prompts | Ran | facebookresearch/three_bricks/main_watermark.py pointer only (licence: NOASSERTION) · get_code("e43d8e6c09e3a48a") |
| load_results | Ran | facebookresearch/three_bricks/main_watermark.py pointer only (licence: NOASSERTION) · get_code("1ab299a3c4b402f6") |
| load_results | Ran | facebookresearch/three_bricks/main_eval.py pointer only (licence: NOASSERTION) · get_code("00e11ed69501d334") |
| ngrams | Ran | this paper's copy was not recorded; identical code first harvested from jwkirchenbauer/lm-watermarking pointer only · get_code("173c14a0656644b7") |
| str2bool | Ran | this paper's copy was not recorded; identical code first harvested from jwkirchenbauer/lm-watermarking pointer only · get_code("678874b625786116") |
Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.
The task of discerning between generated and natural texts is increasingly challenging. In this context, watermarking emerges as a promising technique for ascribing generated text to a specific model. It alters the sampling generation process so as to leave an invisible trace in the generated output, facilitating later detection. This research consolidates watermarks for large language models based on three theoretical and empirical considerations. First, we introduce new statistical tests that offer robust theoretical guarantees which remain valid even at low false-positive rates (less than 10$^{\text{-6}}$). Second, we compare the effectiveness of watermarks using classical benchmarks in the field of natural language processing, gaining insights into their real-world applicability. Third, we develop advanced detection schemes for scenarios where access to the LLM is available, as well as multi-bit watermarking.
The same record, over MCP at https://syntology.ai/mcp:
get_harvested_code_for_paper("2308.00113")
get_code_for_paper("2308.00113")
have("2308.00113")
Connect an agent — have() is free.