# Object Hallucination Detection on CLEVR Source: [Object Hallucination Detection on CLEVR]() Collection index: [llms.txt]() Data updated: 2026-05-13T10:54:39.274Z Evaluation of object hallucination detection performance on the CLEVR dataset, measured using AUROC and AUPR. 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 | Object Hallucination Detection | | Dataset (including any stated split) | CLEVR | | Metric | AUROC | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 35 of 35 | | 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. - [AUPR](): Higher is better; unit: PERCENT. ## Recorded results | Method | AUROC | Result date | Report paper | | --- | --- | --- | --- | | InsLen | 77.72 | Not recorded | 6a0405dfdbc836b87d8278d2 | | InsLen | 75.52 | Not recorded | 6a0405dfdbc836b87d8278d2 | | InsLen | 74.01 | Not recorded | 6a0405dfdbc836b87d8278d2 | | SVAR | 73.01 | Not recorded | 6a0405dfdbc836b87d8278d2 | | Entropy | 72.67 | Not recorded | 6a0405dfdbc836b87d8278d2 | | SVAR | 70.71 | Not recorded | 6a0405dfdbc836b87d8278d2 | | GLSIM | 70.26 | Not recorded | 6a0405dfdbc836b87d8278d2 | | GLSIM | 68.47 | Not recorded | 6a0405dfdbc836b87d8278d2 | | Contextual Lens | 65.62 | Not recorded | 6a0405dfdbc836b87d8278d2 | | NLL | 64.34 | Not recorded | 6a0405dfdbc836b87d8278d2 | | Entropy | 64 | Not recorded | 6a0405dfdbc836b87d8278d2 | | NLL | 63.9 | Not recorded | 6a0405dfdbc836b87d8278d2 | | Entropy | 63.01 | Not recorded | 6a0405dfdbc836b87d8278d2 | | GLSIM | 62.71 | Not recorded | 6a0405dfdbc836b87d8278d2 | | InsLen | 62.34 | Not recorded | 6a0405dfdbc836b87d8278d2 | | SVAR | 62.09 | Not recorded | 6a0405dfdbc836b87d8278d2 | | InsLen | 61.32 | Not recorded | 6a0405dfdbc836b87d8278d2 | | Entropy | 60.93 | Not recorded | 6a0405dfdbc836b87d8278d2 | | NLL | 58.25 | Not recorded | 6a0405dfdbc836b87d8278d2 | | Contextual Lens | 56.89 | Not recorded | 6a0405dfdbc836b87d8278d2 | | Internal Conf. | 56.11 | Not recorded | 6a0405dfdbc836b87d8278d2 | | Internal Conf. | 53.76 | Not recorded | 6a0405dfdbc836b87d8278d2 | | Contextual Lens | 52.39 | Not recorded | 6a0405dfdbc836b87d8278d2 | | Entropy | 51.92 | Not recorded | 6a0405dfdbc836b87d8278d2 | | Internal Conf. | 51.79 | Not recorded | 6a0405dfdbc836b87d8278d2 | | SVAR | 51.23 | Not recorded | 6a0405dfdbc836b87d8278d2 | | SVAR | 50.82 | Not recorded | 6a0405dfdbc836b87d8278d2 | | NLL | 50.58 | Not recorded | 6a0405dfdbc836b87d8278d2 | | GLSIM | 50.31 | Not recorded | 6a0405dfdbc836b87d8278d2 | | NLL | 49.39 | Not recorded | 6a0405dfdbc836b87d8278d2 | | Contextual Lens | 49.32 | Not recorded | 6a0405dfdbc836b87d8278d2 | | GLSIM | 48.61 | Not recorded | 6a0405dfdbc836b87d8278d2 | | Contextual Lens | 47.68 | Not recorded | 6a0405dfdbc836b87d8278d2 | | Internal Conf. | 47.18 | Not recorded | 6a0405dfdbc836b87d8278d2 | | Internal Conf. | 41.72 | Not recorded | 6a0405dfdbc836b87d8278d2 | ### InsLen AUROC: 77.72. Result ID: 6a04580e47f34ce924eaff76. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"Qwen3-VL-8B"}. This record does not assert independent reproduction. ### InsLen AUROC: 75.52. Result ID: 6a04580e47f34ce924eaff75. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"LLaVA-OneVision1.5-8B"}. This record does not assert independent reproduction. ### InsLen AUROC: 74.01. Result ID: 6a04580e47f34ce924eaff74. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"mPLUG-Owl3-8B"}. This record does not assert independent reproduction. ### SVAR AUROC: 73.01. Result ID: 6a04580e47f34ce924eaff67. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"Qwen3-VL-8B"}. This record does not assert independent reproduction. ### Entropy AUROC: 72.67. Result ID: 6a04580e47f34ce924eaff5d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"Qwen3-VL-8B"}. This record does not assert independent reproduction. ### SVAR AUROC: 70.71. Result ID: 6a04580e47f34ce924eaff65. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"mPLUG-Owl3-8B"}. This record does not assert independent reproduction. ### GLSIM AUROC: 70.26. Result ID: 6a04580e47f34ce924eaff70. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"LLaVA-OneVision1.5-8B"}. This record does not assert independent reproduction. ### GLSIM AUROC: 68.47. Result ID: 6a04580e47f34ce924eaff6f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"mPLUG-Owl3-8B"}. This record does not assert independent reproduction. ### Contextual Lens AUROC: 65.62. Result ID: 6a04580e47f34ce924eaff6a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"mPLUG-Owl3-8B"}. This record does not assert independent reproduction. ### NLL AUROC: 64.34. Result ID: 6a04580e47f34ce924eaff56. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"mPLUG-Owl3-8B"}. This record does not assert independent reproduction. ### Entropy AUROC: 64. Result ID: 6a04580e47f34ce924eaff5c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"LLaVA-OneVision1.5-8B"}. This record does not assert independent reproduction. ### NLL AUROC: 63.9. Result ID: 6a04580e47f34ce924eaff57. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"LLaVA-OneVision1.5-8B"}. This record does not assert independent reproduction. ### Entropy AUROC: 63.01. Result ID: 6a04580e47f34ce924eaff5b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"mPLUG-Owl3-8B"}. This record does not assert independent reproduction. ### GLSIM AUROC: 62.71. Result ID: 6a04580e47f34ce924eaff71. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"Qwen3-VL-8B"}. This record does not assert independent reproduction. ### InsLen AUROC: 62.34. Result ID: 6a04580e47f34ce924eaff72. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"LLaVA-1.5-7B"}. This record does not assert independent reproduction. ### SVAR AUROC: 62.09. Result ID: 6a04580e47f34ce924eaff64. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"InstructBLIP-7B"}. This record does not assert independent reproduction. ### InsLen AUROC: 61.32. Result ID: 6a04580e47f34ce924eaff73. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"InstructBLIP-7B"}. This record does not assert independent reproduction. ### Entropy AUROC: 60.93. Result ID: 6a04580e47f34ce924eaff5a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"InstructBLIP-7B"}. This record does not assert independent reproduction. ### NLL AUROC: 58.25. Result ID: 6a04580e47f34ce924eaff58. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"Qwen3-VL-8B"}. This record does not assert independent reproduction. ### Contextual Lens AUROC: 56.89. Result ID: 6a04580e47f34ce924eaff6c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"Qwen3-VL-8B"}. This record does not assert independent reproduction. ### Internal Conf. AUROC: 56.11. Result ID: 6a04580e47f34ce924eaff62. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"Qwen3-VL-8B"}. This record does not assert independent reproduction. ### Internal Conf. AUROC: 53.76. Result ID: 6a04580e47f34ce924eaff60. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"mPLUG-Owl3-8B"}. This record does not assert independent reproduction. ### Contextual Lens AUROC: 52.39. Result ID: 6a04580e47f34ce924eaff6b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"LLaVA-OneVision1.5-8B"}. This record does not assert independent reproduction. ### Entropy AUROC: 51.92. Result ID: 6a04580e47f34ce924eaff59. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"LLaVA-1.5-7B"}. This record does not assert independent reproduction. ### Internal Conf. AUROC: 51.79. Result ID: 6a04580e47f34ce924eaff61. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"LLaVA-OneVision1.5-8B"}. This record does not assert independent reproduction. ### SVAR AUROC: 51.23. Result ID: 6a04580e47f34ce924eaff66. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"LLaVA-OneVision1.5-8B"}. This record does not assert independent reproduction. ### SVAR AUROC: 50.82. Result ID: 6a04580e47f34ce924eaff63. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"LLaVA-1.5-7B"}. This record does not assert independent reproduction. ### NLL AUROC: 50.58. Result ID: 6a04580e47f34ce924eaff54. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"LLaVA-1.5-7B"}. This record does not assert independent reproduction. ### GLSIM AUROC: 50.31. Result ID: 6a04580e47f34ce924eaff6d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"LLaVA-1.5-7B"}. This record does not assert independent reproduction. ### NLL AUROC: 49.39. Result ID: 6a04580e47f34ce924eaff55. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"InstructBLIP-7B"}. This record does not assert independent reproduction. ### Contextual Lens AUROC: 49.32. Result ID: 6a04580e47f34ce924eaff68. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"LLaVA-1.5-7B"}. This record does not assert independent reproduction. ### GLSIM AUROC: 48.61. Result ID: 6a04580e47f34ce924eaff6e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"InstructBLIP-7B"}. This record does not assert independent reproduction. ### Contextual Lens AUROC: 47.68. Result ID: 6a04580e47f34ce924eaff69. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"InstructBLIP-7B"}. This record does not assert independent reproduction. ### Internal Conf. AUROC: 47.18. Result ID: 6a04580e47f34ce924eaff5f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"InstructBLIP-7B"}. This record does not assert independent reproduction. ### Internal Conf. AUROC: 41.72. Result ID: 6a04580e47f34ce924eaff5e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"LLaVA-1.5-7B"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.