# Token-level hallucination detection on HalluRAG Source: [Token-level hallucination detection on HalluRAG]() Collection index: [llms.txt]() Data updated: 2026-07-01T07:11:30.493Z Evaluation of model performance on detecting hallucinations specifically at the token level using the HalluRAG dataset. 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 | Token-level hallucination detection | | Dataset (including any stated split) | HalluRAG | | Metric | AP | | 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 - [AP](): Higher is better; unit: PERCENT. - [AUROC](): Higher is better; unit: PERCENT. ## Recorded results | Method | AP | Result date | Report paper | | --- | --- | --- | --- | | CORTEX | 76.9 | Not recorded | 6a449f02e0224bd4f4992fc1 | | CORTEX | 74.95 | Not recorded | 6a449f02e0224bd4f4992fc1 | | CORTEX | 73.08 | Not recorded | 6a449f02e0224bd4f4992fc1 | | SAPLMA | 68.75 | Not recorded | 6a449f02e0224bd4f4992fc1 | | SAPLMA | 65.84 | Not recorded | 6a449f02e0224bd4f4992fc1 | | RAGLens | 63.8 | Not recorded | 6a449f02e0224bd4f4992fc1 | | SAPLMA | 61.93 | Not recorded | 6a449f02e0224bd4f4992fc1 | | RAGLens | 57.25 | Not recorded | 6a449f02e0224bd4f4992fc1 | | RAGLens | 55.96 | Not recorded | 6a449f02e0224bd4f4992fc1 | | ICR Probe | 32.92 | Not recorded | 6a449f02e0224bd4f4992fc1 | | ICR Probe | 31.97 | Not recorded | 6a449f02e0224bd4f4992fc1 | | ICR Probe | 31.31 | Not recorded | 6a449f02e0224bd4f4992fc1 | | LLM-Check | 29.94 | Not recorded | 6a449f02e0224bd4f4992fc1 | | LLM-Check | 29.44 | Not recorded | 6a449f02e0224bd4f4992fc1 | | LLM-Check | 28.74 | Not recorded | 6a449f02e0224bd4f4992fc1 | | NLL | 26.85 | Not recorded | 6a449f02e0224bd4f4992fc1 | | NLL | 26.58 | Not recorded | 6a449f02e0224bd4f4992fc1 | | NLL | 26.48 | Not recorded | 6a449f02e0224bd4f4992fc1 | ### CORTEX AP: 76.9. Result ID: 6a44bda173d097d845be1b82. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Llama"}. This record does not assert independent reproduction. ### CORTEX AP: 74.95. Result ID: 6a44bda173d097d845be1b83. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen"}. This record does not assert independent reproduction. ### CORTEX AP: 73.08. Result ID: 6a44bda173d097d845be1b84. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Mistral"}. This record does not assert independent reproduction. ### SAPLMA AP: 68.75. Result ID: 6a44bda173d097d845be1b7c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Llama"}. This record does not assert independent reproduction. ### SAPLMA AP: 65.84. Result ID: 6a44bda173d097d845be1b7d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen"}. This record does not assert independent reproduction. ### RAGLens AP: 63.8. Result ID: 6a44bda173d097d845be1b80. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen"}. This record does not assert independent reproduction. ### SAPLMA AP: 61.93. Result ID: 6a44bda173d097d845be1b7e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Mistral"}. This record does not assert independent reproduction. ### RAGLens AP: 57.25. Result ID: 6a44bda173d097d845be1b7f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Llama"}. This record does not assert independent reproduction. ### RAGLens AP: 55.96. Result ID: 6a44bda173d097d845be1b81. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Mistral"}. This record does not assert independent reproduction. ### ICR Probe AP: 32.92. Result ID: 6a44bda173d097d845be1b79. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Llama"}. This record does not assert independent reproduction. ### ICR Probe AP: 31.97. Result ID: 6a44bda173d097d845be1b7b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Mistral"}. This record does not assert independent reproduction. ### ICR Probe AP: 31.31. Result ID: 6a44bda173d097d845be1b7a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen"}. This record does not assert independent reproduction. ### LLM-Check AP: 29.94. Result ID: 6a44bda173d097d845be1b76. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Llama"}. This record does not assert independent reproduction. ### LLM-Check AP: 29.44. Result ID: 6a44bda173d097d845be1b78. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Mistral"}. This record does not assert independent reproduction. ### LLM-Check AP: 28.74. Result ID: 6a44bda173d097d845be1b77. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Qwen"}. This record does not assert independent reproduction. ### NLL AP: 26.85. Result ID: 6a44bda173d097d845be1b75. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Mistral"}. This record does not assert independent reproduction. ### NLL AP: 26.58. Result ID: 6a44bda173d097d845be1b74. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen"}. This record does not assert independent reproduction. ### NLL AP: 26.48. Result ID: 6a44bda173d097d845be1b73. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Llama"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.