List your company
ResearchFormerly WizWand

Language Modeling on HELM macro-averaged (test)

73.2AccuracyZero-Shot Predict
46.78453.64260.567.358Nov 25, 2025
Updated 7mo ago

Evaluation results

MethodLinks
Zero-Shot PredictLanguage Model=o3 Mini2025.1173.2
BFRSLanguage Model=o3 Mini, statistical_significance=true2025.1173.1
MIPROv2Language Model=o3 Mini, statistical_significance=true2025.1173.1
Zero-Shot CoTLanguage Model=o3 Mini, statistical_significance=true2025.1172.7
HELM BaselineLanguage Model=o3 Mini2025.1170.9
MIPROv2Language Model=Claude 3.7 Sonnet, statistical_significance=true2025.1169.8
Zero-Shot CoTLanguage Model=Claude 3.7 Sonnet, statistical_significance=true2025.1169.4
BFRSLanguage Model=Claude 3.7 Sonnet, statistical_significance=true2025.1169.3
Zero-Shot CoTLanguage Model=Gemini 2.0 Flash, statistical_significance=true2025.1166.2
BFRSLanguage Model=Gemini 2.0 Flash, statistical_significance=true2025.1166.2
MIPROv2Language Model=Gemini 2.0 Flash, statistical_significance=true2025.1166.2
BFRSLanguage Model=GPT 4o, statistical_significance=true2025.1165.9
Zero-Shot CoTLanguage Model=GPT 4o, statistical_significance=true2025.1165.7
MIPROv2Language Model=GPT 4o, statistical_significance=true2025.1165.3
Zero-Shot PredictLanguage Model=Claude 3.7 Sonnet2025.1165.1
HELM BaselineLanguage Model=Claude 3.7 Sonnet2025.1164.8
BFRSLanguage Model=Llama 3.3 70B, statistical_significance=true2025.1164.2
Zero-Shot CoTLanguage Model=Llama 3.3 70B, statistical_significance=true2025.1164
MIPROv2Language Model=Llama 3.3 70B, statistical_significance=true2025.1162.1
Zero-Shot PredictLanguage Model=Gemini 2.0 Flash2025.1161.7
HELM BaselineLanguage Model=Gemini 2.0 Flash2025.1161.4
HELM BaselineLanguage Model=GPT 4o2025.1161
MIPROv2Language Model=Qwen3 4B, statistical_significance=true2025.1160.4
BFRSLanguage Model=Qwen3 4B, statistical_significance=true2025.1160.3
Zero-Shot PredictLanguage Model=GPT 4o2025.1159.7
Zero-Shot PredictLanguage Model=Llama 3.3 70B2025.1159.2
Zero-Shot CoTLanguage Model=Qwen3 4B, statistical_significance=true2025.1159
HELM BaselineLanguage Model=Llama 3.3 70B2025.1157.2
Zero-Shot PredictLanguage Model=Qwen3 4B2025.1149.3
HELM BaselineLanguage Model=Qwen3 4B2025.1147.8