Hallucination Detection on HotpotQA In-domain (test)
98.3AccuracyHARP
Evaluation Results
| Method | Links | |
|---|---|---|
| HARPLLM Backbone=Gemma3-12B-it2026.06 | 98.3 | |
| AURORALLM Backbone=Gemma3-12B-it2026.06 | 98.25 | |
| AURORALLM Backbone=Ministral3-8B-Base2026.06 | 98.2 | |
| HARPLLM Backbone=Qwen3-8B-it2026.06 | 98.05 | |
| HARPLLM Backbone=Ministral3-8B-Base2026.06 | 97.95 | |
| ICR probeLLM Backbone=Gemma3-12B-it2026.06 | 97.1 | |
| AURORALLM Backbone=Qwen3-8B-it2026.06 | 97.05 | |
| HARPLLM Backbone=Llama-3-8B-it2026.06 | 96.75 | |
| AURORALLM Backbone=Llama-3-8B-it2026.06 | 96.4 | |
| ICR probeLLM Backbone=Qwen3-8B-it2026.06 | 94.75 | |
| ICR probeLLM Backbone=Llama-3-8B-it2026.06 | 93.55 | |
| ICR probeLLM Backbone=Ministral3-8B-Base2026.06 | 92.8 | |
| InterrogateLLMLLM Backbone=Qwen3-8B-it2026.06 | 68.3 | |
| InterrogateLLMLLM Backbone=Ministral3-8B-Base2026.06 | 67.7 | |
| InterrogateLLMLLM Backbone=Llama-3-8B-it2026.06 | 60.5 | |
| InterrogateLLMLLM Backbone=Gemma3-12B-it2026.06 | 59.5 |