List your company

Tool Use Accuracy on Unseen Tools

1SRtChatHuman
0.487280.620390.75350.88661May 7, 2024

Evaluation results

MethodLinks
ChatHuman2024.0510.9680.8930.9530.953
ChatHumanBackbone=LLaVA-1.5-7B2024.050.9990.9670.8930.9540.953
Visual ChatGPT-4Model Version=GPT-42024.050.9980.9130.8010.8720.907
Ours w/ GPT-4LLM Backbone=GPT-42024.050.9690.9240.7340.7460.876
GPT4Tools-FTFine-tuned=true2024.050.9040.8070.690.7470.8
GPT4ToolsFine-tuned=false2024.050.6120.5460.5420.5250.573
Visual ChatGPT-3.5Model Version=GPT-3.52024.050.5070.3140.2260.2930.803