# Hallucination Detection on HaloQuest Source: [Hallucination Detection on HaloQuest]() Collection index: [llms.txt]() Data updated: 2026-02-26T06:00:19.647Z Evaluation of model performance on hallucination detection using the HaloQuest 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 | Hallucination Detection | | Dataset (including any stated split) | HaloQuest | | Metric | ROC | | Unit | SCALAR | | 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: SCALAR. - [RAC](): Higher is better; unit: SCALAR. - [F1 Score (@B)](): Higher is better; unit: PERCENT. - [RA@50](): Higher is better; unit: PERCENT. - [Sample Latency (ms)](): Lower is better; unit: TIME. - [Inference Latency (s)](): Lower is better; unit: TIME. ## Recorded results | Method | ROC | Result date | Report paper | | --- | --- | --- | --- | | Logistic | 0.834 | Not recorded | 695b458153fdd5c36174310f | | FaithSCAN | 0.829 | Not recorded | 695b458153fdd5c36174310f | | FaithSCAN | 0.827 | Not recorded | 695b458153fdd5c36174310f | | Logistic | 0.795 | Not recorded | 695b458153fdd5c36174310f | | SelfCheckGPT | 0.783 | Not recorded | 695b458153fdd5c36174310f | | Logistic | 0.735 | Not recorded | 695b458153fdd5c36174310f | | P(True) | 0.716 | Not recorded | 695b458153fdd5c36174310f | | FaithSCAN | 0.713 | Not recorded | 695b458153fdd5c36174310f | | SEU | 0.678 | Not recorded | 695b458153fdd5c36174310f | | SEU | 0.66 | Not recorded | 695b458153fdd5c36174310f | | EmbVar | 0.641 | Not recorded | 695b458153fdd5c36174310f | | P(True) | 0.631 | Not recorded | 695b458153fdd5c36174310f | | SE | 0.606 | Not recorded | 695b458153fdd5c36174310f | | EmbVar | 0.605 | Not recorded | 695b458153fdd5c36174310f | | SelfCheckGPT | 0.603 | Not recorded | 695b458153fdd5c36174310f | | SE | 0.595 | Not recorded | 695b458153fdd5c36174310f | | PE | 0.578 | Not recorded | 695b458153fdd5c36174310f | | PE | 0.573 | Not recorded | 695b458153fdd5c36174310f | | T-NLL | 0.571 | Not recorded | 695b458153fdd5c36174310f | | EmbVar | 0.546 | Not recorded | 695b458153fdd5c36174310f | | T-NLL | 0.542 | Not recorded | 695b458153fdd5c36174310f | | SE | 0.537 | Not recorded | 695b458153fdd5c36174310f | | SEU | 0.536 | Not recorded | 695b458153fdd5c36174310f | | P(True) | 0.515 | Not recorded | 695b458153fdd5c36174310f | | SelfCheckGPT | 0.474 | Not recorded | 695b458153fdd5c36174310f | | T-NLL | 0.447 | Not recorded | 695b458153fdd5c36174310f | | PE | 0.44 | Not recorded | 695b458153fdd5c36174310f | ### Logistic ROC: 0.834. Result ID: 6997f5b05698665ec44f33dd. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"LLaVA-8B"}. This record does not assert independent reproduction. ### FaithSCAN ROC: 0.829. Result ID: 6997f5b05698665ec44f33e7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-VL-8B"}. This record does not assert independent reproduction. ### FaithSCAN ROC: 0.827. Result ID: 6997f5b05698665ec44f33de. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"LLaVA-8B"}. This record does not assert independent reproduction. ### Logistic ROC: 0.795. Result ID: 6997f5b05698665ec44f33e6. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-VL-8B"}. This record does not assert independent reproduction. ### SelfCheckGPT ROC: 0.783. Result ID: 6997f5b05698665ec44f33e4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-VL-8B"}. This record does not assert independent reproduction. ### Logistic ROC: 0.735. Result ID: 6997f5b05698665ec44f33d4. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"IB-T5-XL"}. This record does not assert independent reproduction. ### P(True) ROC: 0.716. Result ID: 6997f5b05698665ec44f33dc. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"LLaVA-8B"}. This record does not assert independent reproduction. ### FaithSCAN ROC: 0.713. Result ID: 6997f5b05698665ec44f33d5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"IB-T5-XL"}. This record does not assert independent reproduction. ### SEU ROC: 0.678. Result ID: 6997f5b05698665ec44f33e3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-VL-8B"}. This record does not assert independent reproduction. ### SEU ROC: 0.66. Result ID: 6997f5b05698665ec44f33d1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"IB-T5-XL"}. This record does not assert independent reproduction. ### EmbVar ROC: 0.641. Result ID: 6997f5b05698665ec44f33e1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-VL-8B"}. This record does not assert independent reproduction. ### P(True) ROC: 0.631. Result ID: 6997f5b05698665ec44f33e5. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-VL-8B"}. This record does not assert independent reproduction. ### SE ROC: 0.606. Result ID: 6997f5b05698665ec44f33e2. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-VL-8B"}. This record does not assert independent reproduction. ### EmbVar ROC: 0.605. Result ID: 6997f5b05698665ec44f33cf. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"IB-T5-XL"}. This record does not assert independent reproduction. ### SelfCheckGPT ROC: 0.603. Result ID: 6997f5b05698665ec44f33db. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"LLaVA-8B"}. This record does not assert independent reproduction. ### SE ROC: 0.595. Result ID: 6997f5b05698665ec44f33d9. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"LLaVA-8B"}. This record does not assert independent reproduction. ### PE ROC: 0.578. Result ID: 6997f5b05698665ec44f33cd. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"IB-T5-XL"}. This record does not assert independent reproduction. ### PE ROC: 0.573. Result ID: 6997f5b05698665ec44f33df. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-VL-8B"}. This record does not assert independent reproduction. ### T-NLL ROC: 0.571. Result ID: 6997f5b05698665ec44f33ce. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"IB-T5-XL"}. This record does not assert independent reproduction. ### EmbVar ROC: 0.546. Result ID: 6997f5b05698665ec44f33d8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"LLaVA-8B"}. This record does not assert independent reproduction. ### T-NLL ROC: 0.542. Result ID: 6997f5b05698665ec44f33e0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-VL-8B"}. This record does not assert independent reproduction. ### SE ROC: 0.537. Result ID: 6997f5b05698665ec44f33d0. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"IB-T5-XL"}. This record does not assert independent reproduction. ### SEU ROC: 0.536. Result ID: 6997f5b05698665ec44f33da. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"LLaVA-8B"}. This record does not assert independent reproduction. ### P(True) ROC: 0.515. Result ID: 6997f5b05698665ec44f33d3. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"IB-T5-XL"}. This record does not assert independent reproduction. ### SelfCheckGPT ROC: 0.474. Result ID: 6997f5b05698665ec44f33d2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"IB-T5-XL"}. This record does not assert independent reproduction. ### T-NLL ROC: 0.447. Result ID: 6997f5b05698665ec44f33d7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"LLaVA-8B"}. This record does not assert independent reproduction. ### PE ROC: 0.44. Result ID: 6997f5b05698665ec44f33d6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"LLaVA-8B"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.