# Watermark Detectability on C4 RealNewsLike (Sub-0.5) (test) Source: [Watermark Detectability on C4 RealNewsLike (Sub-0.5) (test)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:25:07.359Z Evaluation of watermark detectability performance using AUC and True Positive Rate (TPR) on the specific test split of the C4 RealNewsLike dataset subjected to a Sub-0.5 perturbation. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Watermark Detectability | | Dataset (including any stated split) | C4 RealNewsLike (Sub-0.5) (test) | | Metric | AUC | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 14 of 14 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [AUC](): Higher is better; unit: PERCENT. - [TPR](): Higher is better; unit: PERCENT. ## Recorded results | Method | AUC | Result date | Report paper | | --- | --- | --- | --- | | Gumbel-max | 86.2 | Not recorded | 693115d9d6b0dd3342933da6 | | SynthID | 84.3 | Not recorded | 693115d9d6b0dd3342933da6 | | Gumbel-max | 84 | Not recorded | 693115d9d6b0dd3342933da6 | | KGW-D | 83.7 | Not recorded | 693115d9d6b0dd3342933da6 | | SynthID | 83.4 | Not recorded | 693115d9d6b0dd3342933da6 | | MarkTune | 80.9 | Not recorded | 693115d9d6b0dd3342933da6 | | MarkTune | 78.3 | Not recorded | 693115d9d6b0dd3342933da6 | | GaussMark | 77.5 | Not recorded | 693115d9d6b0dd3342933da6 | | KGW-2 | 77.2 | Not recorded | 693115d9d6b0dd3342933da6 | | KGW-D | 75.9 | Not recorded | 693115d9d6b0dd3342933da6 | | GaussMark | 75.9 | Not recorded | 693115d9d6b0dd3342933da6 | | KGW-2 | 74.7 | Not recorded | 693115d9d6b0dd3342933da6 | | KGW-1 | 63.8 | Not recorded | 693115d9d6b0dd3342933da6 | | KGW-1 | 63.8 | Not recorded | 693115d9d6b0dd3342933da6 | ### Gumbel-max AUC: 86.2. Result ID: 69998b265698665ec4fd0d38. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Llama2-7B","Temperature":"1.0"}. This record does not assert independent reproduction. ### SynthID AUC: 84.3. Result ID: 69998b265698665ec4fd0d39. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Llama2-7B","Temperature":"1.0"}. This record does not assert independent reproduction. ### Gumbel-max AUC: 84. Result ID: 69998b265698665ec4fd0d31. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-4B","Temperature":"0.7"}. This record does not assert independent reproduction. ### KGW-D AUC: 83.7. Result ID: 69998b265698665ec4fd0d33. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-4B","Temperature":"0.7"}. This record does not assert independent reproduction. ### SynthID AUC: 83.4. Result ID: 69998b265698665ec4fd0d32. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-4B","Temperature":"0.7"}. This record does not assert independent reproduction. ### MarkTune AUC: 80.9. Result ID: 69998b265698665ec4fd0d3c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Llama2-7B","Temperature":"1.0"}. This record does not assert independent reproduction. ### MarkTune AUC: 78.3. Result ID: 69998b265698665ec4fd0d35. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-4B","Temperature":"0.7"}. This record does not assert independent reproduction. ### GaussMark AUC: 77.5. Result ID: 69998b265698665ec4fd0d34. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-4B","Temperature":"0.7"}. This record does not assert independent reproduction. ### KGW-2 AUC: 77.2. Result ID: 69998b265698665ec4fd0d37. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Llama2-7B","Temperature":"1.0"}. This record does not assert independent reproduction. ### KGW-D AUC: 75.9. Result ID: 69998b265698665ec4fd0d3a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Llama2-7B","Temperature":"1.0"}. This record does not assert independent reproduction. ### GaussMark AUC: 75.9. Result ID: 69998b265698665ec4fd0d3b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Llama2-7B","Temperature":"1.0"}. This record does not assert independent reproduction. ### KGW-2 AUC: 74.7. Result ID: 69998b265698665ec4fd0d30. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-4B","Temperature":"0.7"}. This record does not assert independent reproduction. ### KGW-1 AUC: 63.8. Result ID: 69998b265698665ec4fd0d2f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-4B","Temperature":"0.7"}. This record does not assert independent reproduction. ### KGW-1 AUC: 63.8. Result ID: 69998b265698665ec4fd0d36. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Llama2-7B","Temperature":"1.0"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.