List your company
ResearchFormerly WizWand

Safety Evaluation on BeaverTails Evaluation

1.24Harmful Score (HS)Antibody
-0.94813.82128.5943.359Feb 28, 2026
Updated 8mo ago

Evaluation results

MethodLinks
AntibodyBackbone=Llama-2-7B, Fine-tuning Dataset=AGNEWS2026.021.24
AntibodyBackbone=Llama-2-7B, Fine-tuning Dataset=GSM8K2026.021.24
AntibodyBackbone=Llama-2-7B, Fine-tuning Dataset=SST22026.021.48
LisaBackbone=Llama-2-7B, Fine-tuning Dataset=GSM8K2026.025.86
BoosterBackbone=Llama-2-7B, Fine-tuning Dataset=GSM8K2026.029.06
LisaBackbone=Llama-2-7B, Fine-tuning Dataset=AlpacaEval2026.0211.44
BoosterBackbone=Llama-2-7B, Fine-tuning Dataset=SST22026.0214.31
BoosterBackbone=Llama-2-7B, Fine-tuning Dataset=AGNEWS2026.0215.88
LisaBackbone=Llama-2-7B, Fine-tuning Dataset=AGNEWS2026.0220.93
LisaBackbone=Llama-2-7B, Fine-tuning Dataset=SST22026.0222.94
VaccineBackbone=Llama-2-7B, Fine-tuning Dataset=GSM8K2026.0223.6
SFTBackbone=Llama-2-7B, Fine-tuning Dataset=GSM8K2026.0223.94
AntibodyBackbone=Llama-2-7B, Fine-tuning Dataset=AlpacaEval2026.0224.18
SFTBackbone=Llama-2-7B, Fine-tuning Dataset=AGNEWS2026.0234.57
SFTBackbone=Llama-2-7B, Fine-tuning Dataset=SST22026.0236.29
BoosterBackbone=Llama-2-7B, Fine-tuning Dataset=AlpacaEval2026.0236.91
SFTBackbone=Llama-2-7B, Fine-tuning Dataset=AlpacaEval2026.0239.48
VaccineBackbone=Llama-2-7B, Fine-tuning Dataset=AGNEWS2026.0239.73
VaccineBackbone=Llama-2-7B, Fine-tuning Dataset=SST22026.0244.16
VaccineBackbone=Llama-2-7B, Fine-tuning Dataset=AlpacaEval2026.0255.94