Hallucination Detection on TriviaQA In-domain (test)
99.1AccuracyAURORA
Evaluation Results
| Method | Links | |
|---|---|---|
| AURORALLM Backbone=Gemma3-12B-it2026.06 | 99.1 | |
| HARPLLM Backbone=Ministral3-8B-Base2026.06 | 99.05 | |
| HARPLLM Backbone=Qwen3-8B-it2026.06 | 99 | |
| HARPLLM Backbone=Gemma3-12B-it2026.06 | 98.65 | |
| AURORALLM Backbone=Ministral3-8B-Base2026.06 | 98.45 | |
| AURORALLM Backbone=Llama-3-8B-it2026.06 | 98.25 | |
| HARPLLM Backbone=Llama-3-8B-it2026.06 | 98.2 | |
| AURORALLM Backbone=Qwen3-8B-it2026.06 | 97.9 | |
| ICR probeLLM Backbone=Gemma3-12B-it2026.06 | 97.65 | |
| ICR probeLLM Backbone=Qwen3-8B-it2026.06 | 95 | |
| ICR probeLLM Backbone=Llama-3-8B-it2026.06 | 93.75 | |
| ICR probeLLM Backbone=Ministral3-8B-Base2026.06 | 93.4 | |
| InterrogateLLMLLM Backbone=Ministral3-8B-Base2026.06 | 84.1 | |
| InterrogateLLMLLM Backbone=Qwen3-8B-it2026.06 | 76.45 | |
| InterrogateLLMLLM Backbone=Gemma3-12B-it2026.06 | 69.05 | |
| InterrogateLLMLLM Backbone=Llama-3-8B-it2026.06 | 57.3 |