List your company

Hallucination Detection on Summedits (test)

93AUCSeparate Guardrail 9B
57.6466.827685.18May 30, 2025

Evaluation results

MethodLinks
Separate Guardrail 9BBase model=Gemma2-9B2025.0593
Separate Guardrail 7BBase model=Mistral-7B-v0.32025.0592
DSA:LSTBase model=Mistral-7B-v0.32025.0590
DSA:LSTBase model=Gemma2-9B2025.0589
DSA:LSTBase model=Qwen2.5-7B2025.0588
Separate Guardrail 7BBase model=Qwen2.5-7B2025.0588
DSA:LSTBase model=Qwen2.5-14B2025.0588
Separate Guardrail 14BBase model=Qwen2.5-14B2025.0588
DSA:PLRBase model=Qwen2.5-7B2025.0583
DSA:RTBBase model=Qwen2.5-7B2025.0583
DSA:RTBBase model=Qwen2.5-14B2025.0583
DSA:RTBBase model=Gemma2-9B2025.0583
DSA:PLRBase model=Qwen2.5-14B2025.0581
Suffix-ZSPromptBase model=Gemma2-9B2025.0578
DSA:RTBBase model=Mistral-7B-v0.32025.0572
Suffix-ZSPromptBase model=Qwen2.5-14B2025.0571
DSA:PLRBase model=Gemma2-9B2025.0570
Suffix-ZSPromptBase model=Qwen2.5-7B2025.0564
Separate Guardrail 51MBase model=Qwen2.5-14B2025.0562
Suffix-ZSPromptBase model=Mistral-7B-v0.32025.0562
DSA:PLRBase model=Mistral-7B-v0.32025.0562
Separate Guardrail 25MBase model=Qwen2.5-7B2025.0561
Separate Guardrail 27MBase model=Mistral-7B-v0.32025.0559
Separate Guardrail 31MBase model=Gemma2-9B2025.0559