List your company

Hallucination Detection on HotpotQA In-domain (test)

98.3AccuracyHARP
57.94868.42478.989.376Jun 28, 2026

Evaluation results

MethodLinks
HARPLLM Backbone=Gemma3-12B-it2026.0698.3
AURORALLM Backbone=Gemma3-12B-it2026.0698.25
AURORALLM Backbone=Ministral3-8B-Base2026.0698.2
HARPLLM Backbone=Qwen3-8B-it2026.0698.05
HARPLLM Backbone=Ministral3-8B-Base2026.0697.95
ICR probeLLM Backbone=Gemma3-12B-it2026.0697.1
AURORALLM Backbone=Qwen3-8B-it2026.0697.05
HARPLLM Backbone=Llama-3-8B-it2026.0696.75
AURORALLM Backbone=Llama-3-8B-it2026.0696.4
ICR probeLLM Backbone=Qwen3-8B-it2026.0694.75
ICR probeLLM Backbone=Llama-3-8B-it2026.0693.55
ICR probeLLM Backbone=Ministral3-8B-Base2026.0692.8
InterrogateLLMLLM Backbone=Qwen3-8B-it2026.0668.3
InterrogateLLMLLM Backbone=Ministral3-8B-Base2026.0667.7
InterrogateLLMLLM Backbone=Llama-3-8B-it2026.0660.5
InterrogateLLMLLM Backbone=Gemma3-12B-it2026.0659.5