List your company

Mathematical Reasoning on GSM8K (100 examples)

5.47Speed-upEAGLE-3
1.29962.38233.4654.5477May 13, 2026

Evaluation results

MethodLinks
EAGLE-3Model=Vicuna-13B, Attack Status=Clean2026.055.475.95——
EAGLE-2Model=Vicuna-7B, Attack Status=Clean2026.054.994.68——
EAGLE-2Model=Vicuna-13B, Attack Status=Clean2026.054.844.91——
HydraModel=Vicuna-7B, Attack Status=Clean2026.054.373.84——
HydraModel=Vicuna-13B, Attack Status=Clean2026.054.293.58——
EAGLEModel=Vicuna-7B, Attack Status=Clean2026.054.113.62——
EAGLEModel=Vicuna-13B, Attack Status=Clean2026.053.663.79——
MedusaModel=Vicuna-7B, Attack Status=Clean2026.053.462.47——
MedusaModel=Vicuna-13B, Attack Status=Clean2026.053.392.51——
EAGLE-2Model=Vicuna-7B, Attack Status=Attacked (MISTLETOE)2026.052.712.84——
EAGLE-2Model=Vicuna-13B, Attack Status=Attacked (MISTLETOE)2026.052.623.25——
MedusaModel=Vicuna-7B, Attack Status=Attacked (MISTLETOE)2026.052.51.82——
HydraModel=Vicuna-7B, Attack Status=Attacked (MISTLETOE)2026.052.243.07——
EAGLEModel=Vicuna-7B, Attack Status=Attacked (MISTLETOE)2026.052.023.06——
EAGLE-3Model=Vicuna-13B, Attack Status=Attacked (MISTLETOE)2026.051.832.79——
HydraModel=Vicuna-13B, Attack Status=Attacked (MISTLETOE)2026.051.723.07——
EAGLEModel=Vicuna-13B, Attack Status=Attacked (MISTLETOE)2026.051.692.96——
MedusaModel=Vicuna-13B, Attack Status=Attacked (MISTLETOE)2026.051.462.32——