# Hallucination Detection on HalLoc-VQA Source: [Hallucination Detection on HalLoc-VQA]() Collection index: [llms.txt]() Data updated: 2026-02-26T06:00:19.630Z Evaluation of model performance on general hallucination detection using the HalLoc-VQA dataset, reporting metrics such as ROC, RAC, F1@B, RA@50, and inference latency. 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 | Hallucination Detection | | Dataset (including any stated split) | HalLoc-VQA | | Metric | ROC | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 27 of 27 | | 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 - [ROC](): Higher is better; unit: PERCENT. - [RAC](): Higher is better; unit: PERCENT. - [F1@B](): Higher is better; unit: PERCENT. - [RA@50](): Higher is better; unit: PERCENT. - [Latency (Sample)](): Lower is better; unit: TIME. - [Latency (Inference)](): Lower is better; unit: TIME. ## Recorded results | Method | ROC | Result date | Report paper | | --- | --- | --- | --- | | FaithSCAN | 77.9 | Not recorded | 695b458153fdd5c36174310f | | SEU | 76.9 | Not recorded | 695b458153fdd5c36174310f | | FaithSCAN | 75.4 | Not recorded | 695b458153fdd5c36174310f | | P(True) | 75 | Not recorded | 695b458153fdd5c36174310f | | EmbVar | 74.7 | Not recorded | 695b458153fdd5c36174310f | | FaithSCAN | 73.9 | Not recorded | 695b458153fdd5c36174310f | | Logistic | 73.3 | Not recorded | 695b458153fdd5c36174310f | | SelfCheckGPT | 71.5 | Not recorded | 695b458153fdd5c36174310f | | PE | 68.6 | Not recorded | 695b458153fdd5c36174310f | | SE | 68.3 | Not recorded | 695b458153fdd5c36174310f | | P(True) | 67.5 | Not recorded | 695b458153fdd5c36174310f | | Logistic | 67.2 | Not recorded | 695b458153fdd5c36174310f | | Logistic | 66.9 | Not recorded | 695b458153fdd5c36174310f | | SEU | 66.3 | Not recorded | 695b458153fdd5c36174310f | | EmbVar | 65.7 | Not recorded | 695b458153fdd5c36174310f | | SelfCheckGPT | 65.6 | Not recorded | 695b458153fdd5c36174310f | | SEU | 65.3 | Not recorded | 695b458153fdd5c36174310f | | T-NLL | 65 | Not recorded | 695b458153fdd5c36174310f | | EmbVar | 62.8 | Not recorded | 695b458153fdd5c36174310f | | P(True) | 61.5 | Not recorded | 695b458153fdd5c36174310f | | T-NLL | 60.7 | Not recorded | 695b458153fdd5c36174310f | | PE | 60.5 | Not recorded | 695b458153fdd5c36174310f | | SE | 58.3 | Not recorded | 695b458153fdd5c36174310f | | SE | 55.9 | Not recorded | 695b458153fdd5c36174310f | | PE | 53.6 | Not recorded | 695b458153fdd5c36174310f | | T-NLL | 53.5 | Not recorded | 695b458153fdd5c36174310f | | SelfCheckGPT | 50.9 | Not recorded | 695b458153fdd5c36174310f | ### FaithSCAN ROC: 77.9. Result ID: 6997f5b05698665ec44f33c3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"LLaVA-8B"}. This record does not assert independent reproduction. ### SEU ROC: 76.9. Result ID: 6997f5b05698665ec44f33bf. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"LLaVA-8B"}. This record does not assert independent reproduction. ### FaithSCAN ROC: 75.4. Result ID: 6997f5b05698665ec44f33ba. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"IB-T5-XL"}. This record does not assert independent reproduction. ### P(True) ROC: 75. Result ID: 6997f5b05698665ec44f33ca. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-VL-8B"}. This record does not assert independent reproduction. ### EmbVar ROC: 74.7. Result ID: 6997f5b05698665ec44f33bd. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"LLaVA-8B"}. This record does not assert independent reproduction. ### FaithSCAN ROC: 73.9. Result ID: 6997f5b05698665ec44f33cc. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-VL-8B"}. This record does not assert independent reproduction. ### Logistic ROC: 73.3. Result ID: 6997f5b05698665ec44f33c2. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"LLaVA-8B"}. This record does not assert independent reproduction. ### SelfCheckGPT ROC: 71.5. Result ID: 6997f5b05698665ec44f33c0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"LLaVA-8B"}. This record does not assert independent reproduction. ### PE ROC: 68.6. Result ID: 6997f5b05698665ec44f33c4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-VL-8B"}. This record does not assert independent reproduction. ### SE ROC: 68.3. Result ID: 6997f5b05698665ec44f33be. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"LLaVA-8B"}. This record does not assert independent reproduction. ### P(True) ROC: 67.5. Result ID: 6997f5b05698665ec44f33b8. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"IB-T5-XL"}. This record does not assert independent reproduction. ### Logistic ROC: 67.2. Result ID: 6997f5b05698665ec44f33cb. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-VL-8B"}. This record does not assert independent reproduction. ### Logistic ROC: 66.9. Result ID: 6997f5b05698665ec44f33b9. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"IB-T5-XL"}. This record does not assert independent reproduction. ### SEU ROC: 66.3. Result ID: 6997f5b05698665ec44f33b6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"IB-T5-XL"}. This record does not assert independent reproduction. ### EmbVar ROC: 65.7. Result ID: 6997f5b05698665ec44f33b4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"IB-T5-XL"}. This record does not assert independent reproduction. ### SelfCheckGPT ROC: 65.6. Result ID: 6997f5b05698665ec44f33c9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-VL-8B"}. This record does not assert independent reproduction. ### SEU ROC: 65.3. Result ID: 6997f5b05698665ec44f33c8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-VL-8B"}. This record does not assert independent reproduction. ### T-NLL ROC: 65. Result ID: 6997f5b05698665ec44f33c5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-VL-8B"}. This record does not assert independent reproduction. ### EmbVar ROC: 62.8. Result ID: 6997f5b05698665ec44f33c6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-VL-8B"}. This record does not assert independent reproduction. ### P(True) ROC: 61.5. Result ID: 6997f5b05698665ec44f33c1. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"LLaVA-8B"}. This record does not assert independent reproduction. ### T-NLL ROC: 60.7. Result ID: 6997f5b05698665ec44f33bc. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"LLaVA-8B"}. This record does not assert independent reproduction. ### PE ROC: 60.5. Result ID: 6997f5b05698665ec44f33bb. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"LLaVA-8B"}. This record does not assert independent reproduction. ### SE ROC: 58.3. Result ID: 6997f5b05698665ec44f33b5. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"IB-T5-XL"}. This record does not assert independent reproduction. ### SE ROC: 55.9. Result ID: 6997f5b05698665ec44f33c7. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-VL-8B"}. This record does not assert independent reproduction. ### PE ROC: 53.6. Result ID: 6997f5b05698665ec44f33b2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"IB-T5-XL"}. This record does not assert independent reproduction. ### T-NLL ROC: 53.5. Result ID: 6997f5b05698665ec44f33b3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"IB-T5-XL"}. This record does not assert independent reproduction. ### SelfCheckGPT ROC: 50.9. Result ID: 6997f5b05698665ec44f33b7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"IB-T5-XL"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.