List your company

Refusal Rate Evaluation on FBA (test)

13Refusal RateDEEPSEEK-R1-DISTILL-LLAMA-8B
9.6432.325577.68May 11, 2026

Evaluation results

MethodLinks
DEEPSEEK-R1-DISTILL-LLAMA-8BOnline Alignment=Offline Only, Offline SAT Variant=SafeDPO SAT2026.0513
DEEPSEEK-R1-DISTILL-LLAMA-8BOnline Alignment=Offline Only, Offline SAT Variant=Base2026.0514
LLAMA-3.2-1B-INSTRUCTOnline Alignment=Offline Only, Offline SAT Variant=Base2026.0515
DEEPSEEK-R1-DISTILL-LLAMA-8BOnline Alignment=Offline Only, Offline SAT Variant=DPO SAT2026.0515
DEEPSEEK-R1-DISTILL-QWEN-14BOnline Alignment=Offline Only, Offline SAT Variant=Base2026.0516
DEEPSEEK-R1-DISTILL-QWEN-14BOnline Alignment=Offline Only, Offline SAT Variant=DPO SAT2026.0518
DEEPSEEK-R1-DISTILL-QWEN-14BOnline Alignment=Offline Only, Offline SAT Variant=SafeDPO SAT2026.0519
LLAMA-3.2-1B-INSTRUCTOnline Alignment=RAG-Pref, Offline SAT Variant=Base2026.0528
LLAMA-3.2-1B-INSTRUCTOnline Alignment=Offline Only, Offline SAT Variant=DPO SAT2026.0531
Gemma-2-2B-ITOnline Alignment=Offline Only, Offline SAT Variant=Base2026.0532
LLAMA-3.1-8B-INSTRUCTOnline Alignment=Offline Only, Offline SAT Variant=Base2026.0535
LLAMA-3.1-8B-INSTRUCTOnline Alignment=OPAD, Offline SAT Variant=SafeDPO SAT2026.0537
LLAMA-3.2-1B-INSTRUCTOnline Alignment=Offline Only, Offline SAT Variant=SafeDPO SAT2026.0540
DEEPSEEK-R1-DISTILL-QWEN-14BOnline Alignment=OPAD, Offline SAT Variant=Base2026.0541
LLAMA-3.1-8B-INSTRUCTOnline Alignment=Offline Only, Offline SAT Variant=DPO SAT2026.0543
LLAMA-3.1-8B-INSTRUCTOnline Alignment=OPAD, Offline SAT Variant=Base2026.0543
DEEPSEEK-R1-DISTILL-LLAMA-8BOnline Alignment=OPAD, Offline SAT Variant=Base2026.0544
DEEPSEEK-R1-DISTILL-QWEN-14BOnline Alignment=OPAD, Offline SAT Variant=DPO SAT2026.0544
Gemma-2-2B-ITOnline Alignment=Offline Only, Offline SAT Variant=DPO SAT2026.0545
LLAMA-3.1-8B-INSTRUCTOnline Alignment=Offline Only, Offline SAT Variant=SafeDPO SAT2026.0545
DEEPSEEK-R1-DISTILL-LLAMA-8BOnline Alignment=OPAD, Offline SAT Variant=SafeDPO SAT2026.0545
DEEPSEEK-R1-DISTILL-QWEN-14BOnline Alignment=OPAD, Offline SAT Variant=SafeDPO SAT2026.0546
Gemma-2-2B-ITOnline Alignment=Offline Only, Offline SAT Variant=SafeDPO SAT2026.0547
LLAMA-3.1-8B-INSTRUCTOnline Alignment=OPAD, Offline SAT Variant=DPO SAT2026.0547
DEEPSEEK-R1-DISTILL-LLAMA-8BOnline Alignment=OPAD, Offline SAT Variant=DPO SAT2026.0547
LLAMA-3.2-1B-INSTRUCTOnline Alignment=RAG-Pref, Offline SAT Variant=DPO SAT2026.0558
LLAMA-3.2-1B-INSTRUCTOnline Alignment=OPAD, Offline SAT Variant=Base2026.0559
DEEPSEEK-R1-DISTILL-LLAMA-8BOnline Alignment=RAG-Pref, Offline SAT Variant=Base2026.0559
DEEPSEEK-R1-DISTILL-LLAMA-8BOnline Alignment=RAG-Pref, Offline SAT Variant=DPO SAT2026.0559
DEEPSEEK-R1-DISTILL-LLAMA-8BOnline Alignment=RAG-Pref, Offline SAT Variant=SafeDPO SAT2026.0559
LLAMA-3.2-1B-INSTRUCTOnline Alignment=OPAD, Offline SAT Variant=DPO SAT2026.0561
Gemma-2-2B-ITOnline Alignment=RAG-Pref, Offline SAT Variant=Base2026.0563
DEEPSEEK-R1-DISTILL-QWEN-14BOnline Alignment=RAG-Pref, Offline SAT Variant=Base2026.0564
DEEPSEEK-R1-DISTILL-QWEN-14BOnline Alignment=RAG-Pref, Offline SAT Variant=SafeDPO SAT2026.0564
LLAMA-3.2-1B-INSTRUCTOnline Alignment=OPAD, Offline SAT Variant=SafeDPO SAT2026.0566
DEEPSEEK-R1-DISTILL-QWEN-14BOnline Alignment=RAG-Pref, Offline SAT Variant=DPO SAT2026.0568
Gemma-2-2B-ITOnline Alignment=RAG-Pref, Offline SAT Variant=DPO SAT2026.0574
Gemma-2-2B-ITOnline Alignment=RAG-Pref, Offline SAT Variant=SafeDPO SAT2026.0575
LLAMA-3.2-1B-INSTRUCTOnline Alignment=RAG-Pref, Offline SAT Variant=SafeDPO SAT2026.0588
LLAMA-3.1-8B-INSTRUCTOnline Alignment=RAG-Pref, Offline SAT Variant=Base2026.0595
LLAMA-3.1-8B-INSTRUCTOnline Alignment=RAG-Pref, Offline SAT Variant=DPO SAT2026.0597
LLAMA-3.1-8B-INSTRUCTOnline Alignment=RAG-Pref, Offline SAT Variant=SafeDPO SAT2026.0597