HyperAI

Conversational Web Navigation On Weblinx

Metrics

Element (IoU)
Intent Match
Overall score
Text (F1)

Results

Performance results of various models on this benchmark

Model Name
Element (IoU)
Intent Match
Overall score
Text (F1)
Paper TitleRepository
GPT-3.5T (Zero-Shot)8.6242.778.513.45--
S-LLaMA-1.3B20.5483.3223.7325.85--
Pix2Act-1.3B8.2881.8016.8825.21--
MindAct-3B16.5079.8920.9423.16--
Fuyu-8B15.7080.0719.9722.30--
Llama-2-13B22.8281.9125.2126.60--
GPT-3.5F18.6477.5621.2222.39--
MindAct-780M13.3975.8715.1313.58--
Flan-T5-780M15.3680.0217.2714.05--
MindAct-250M12.0574.2512.637.67--
Pix2Act-282M6.2079.7112.5116.40--
S-LLaMA-2.7B22.6084.0025.0227.17--
GPT-4T (Zero-Shot)10.8541.6610.726.75--
Flan-T5-250M14.8679.6914.999.21--
Flan-T5-3B20.3181.1423.7725.75--
GPT-4V (Zero-Shot)10.9142.3610.456.21--
Llama-2-7B22.2682.6424.5726.50--
0 of 17 row(s) selected.