List your company

Human Intervention Prediction on COWCORPUS (All Data)

85.3Step AccuracyAlways No Interv
11.87630.9385069.062Feb 19, 2026

Evaluation results

MethodLinks
Always No Intervtype=Baseline2026.0285.3085.301000920
Gemma 27B (finetuned)shots=0, reasoning=false2026.0285.35087.721.696.330.291.830.3
Llava 8B (finetuned)shots=0, reasoning=false2026.0281.747.187.621.692.129.689.720.1
Claude 3.5 Sonnetshots=2, reasoning=false2026.0274.915.885.416.2851685.214.9
GPT-4oshots=0, reasoning=false2026.0274.118.28621.683.219.884.614.7
Claude 3.5 Sonnetshots=2, reasoning=true2026.0272.114.985.318.981.316.783.315.8
Claude 3.5 Sonnetshots=0, reasoning=true2026.0269.716.985.92777.120.881.316.4
Gemini 1.5 Proshots=0, reasoning=true2026.0268.921.187.840.573.827.880.223.7
Claude 3.5 Sonnetshots=0, reasoning=false2026.0268.117.986.432.474.323.179.929.3
Gemini 1.5 Proshots=0, reasoning=false2026.0268.121.388.143.272.428.679.526.2
GPT-4oshots=2, reasoning=false2026.0266.114.785.22772.91978.620.6
Gemini 1.5 Proshots=2, reasoning=true2026.0264.122.189.756.865.431.875.724.3
Gemini 1.5 Proshots=2, reasoning=false2026.0258.618.187.751.459.826.871.128.7
Gemma 27B (base)shots=0, reasoning=false2026.0223.915.489.791.912.126.421.418.7
Llava 8B (base)shots=0, reasoning=false2026.0218.3085.2021.5034.31.7
Always Intervtype=Baseline2026.0214.714.70100025.7015.1