# Safety Detection on Polyguard Education Source: [Safety Detection on Polyguard Education]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:26:05.344Z Evaluation of model performance on Safety Detection using the Polyguard Education dataset, reporting AUROC and FPR@95. 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 | Safety Detection | | Dataset (including any stated split) | Polyguard Education | | Metric | AUROC | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 18 of 18 | | 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 - [AUROC](): Higher is better; unit: PERCENT. - [FPR@95](): Lower is better; unit: PERCENT. ## Recorded results | Method | AUROC | Result date | Report paper | | --- | --- | --- | --- | | T3 | 99.43 | Not recorded | 6984242aa8b01e76cc8ee806 | | T3 | 99.13 | Not recorded | 6984242aa8b01e76cc8ee806 | | LlamaGuard3 | 86.61 | Not recorded | 6984242aa8b01e76cc8ee806 | | CIDER | 74.06 | Not recorded | 6984242aa8b01e76cc8ee806 | | MDJudge | 69.09 | Not recorded | 6984242aa8b01e76cc8ee806 | | WildGuard | 66.37 | Not recorded | 6984242aa8b01e76cc8ee806 | | DuoGuard | 66.26 | Not recorded | 6984242aa8b01e76cc8ee806 | | RMD | 65.05 | Not recorded | 6984242aa8b01e76cc8ee806 | | GMM | 62.88 | Not recorded | 6984242aa8b01e76cc8ee806 | | FDBD | 58.71 | Not recorded | 6984242aa8b01e76cc8ee806 | | LlamaGuard3 | 57.4 | Not recorded | 6984242aa8b01e76cc8ee806 | | LlamaGuard4 | 53.89 | Not recorded | 6984242aa8b01e76cc8ee806 | | NNGUIDE | 50.38 | Not recorded | 6984242aa8b01e76cc8ee806 | | VIM | 50.22 | Not recorded | 6984242aa8b01e76cc8ee806 | | OPENMAX | 47.12 | Not recorded | 6984242aa8b01e76cc8ee806 | | REACT | 44.78 | Not recorded | 6984242aa8b01e76cc8ee806 | | Polyguard | 44.64 | Not recorded | 6984242aa8b01e76cc8ee806 | | ADASCALE | 44.59 | Not recorded | 6984242aa8b01e76cc8ee806 | ### T3 AUROC: 99.43. Result ID: 69964c015698665ec4906163. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"classifier":"OCSVM"}. This record does not assert independent reproduction. ### T3 AUROC: 99.13. Result ID: 69964c015698665ec4906162. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"classifier":"GMM"}. This record does not assert independent reproduction. ### LlamaGuard3 AUROC: 86.61. Result ID: 69964c015698665ec490615e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"parameters":"1B","output\_format":"logits"}. This record does not assert independent reproduction. ### CIDER AUROC: 74.06. Result ID: 69964c015698665ec4906153. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### MDJudge AUROC: 69.09. Result ID: 69964c015698665ec490615f. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### WildGuard AUROC: 66.37. Result ID: 69964c015698665ec490615d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### DuoGuard AUROC: 66.26. Result ID: 69964c015698665ec4906160. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### RMD AUROC: 65.05. Result ID: 69964c015698665ec4906159. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### GMM AUROC: 62.88. Result ID: 69964c015698665ec4906155. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### FDBD AUROC: 58.71. Result ID: 69964c015698665ec4906154. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### LlamaGuard3 AUROC: 57.4. Result ID: 69964c015698665ec490615b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"parameters":"1B"}. This record does not assert independent reproduction. ### LlamaGuard4 AUROC: 53.89. Result ID: 69964c015698665ec490615c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"parameters":"12B"}. This record does not assert independent reproduction. ### NNGUIDE AUROC: 50.38. Result ID: 69964c015698665ec4906156. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### VIM AUROC: 50.22. Result ID: 69964c015698665ec490615a. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### OPENMAX AUROC: 47.12. Result ID: 69964c015698665ec4906157. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### REACT AUROC: 44.78. Result ID: 69964c015698665ec4906158. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### Polyguard AUROC: 44.64. Result ID: 69964c015698665ec4906161. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### ADASCALE AUROC: 44.59. Result ID: 69964c015698665ec4906152. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.