List your company

Hallucination Detection on SQuAD In-domain (test)

92.2AccuracyAURORA
63.0870.6478.285.76Jun 28, 2026

Evaluation results

MethodLinks
AURORALLM Backbone=Gemma3-12B-it2026.0692.2
HARPLLM Backbone=Gemma3-12B-it2026.0691.95
HARPLLM Backbone=Ministral3-8B-Base2026.0689.25
HARPLLM Backbone=Llama-3-8B-it2026.0687.45
HARPLLM Backbone=Qwen3-8B-it2026.0687.05
AURORALLM Backbone=Llama-3-8B-it2026.0685.2
AURORALLM Backbone=Qwen3-8B-it2026.0685.2
AURORALLM Backbone=Ministral3-8B-Base2026.0685.05
ICR probeLLM Backbone=Llama-3-8B-it2026.0678.8
ICR probeLLM Backbone=Ministral3-8B-Base2026.0676.75
ICR probeLLM Backbone=Gemma3-12B-it2026.0673.25
InterrogateLLMLLM Backbone=Qwen3-8B-it2026.0672.8
InterrogateLLMLLM Backbone=Llama-3-8B-it2026.0669.6
ICR probeLLM Backbone=Qwen3-8B-it2026.0668.15
InterrogateLLMLLM Backbone=Gemma3-12B-it2026.0667.2
InterrogateLLMLLM Backbone=Ministral3-8B-Base2026.0664.2