• Model Purpose: Simplified Demo
• GPT Model: gpt-4-turbo-preview
• Training Dataset: Simplified
• Accuracy: 92.3%
• Best response in list (abstract prompt): 71.9%
• Best response in list (precise prompt): 97.6%
• Needle In A Haystack: 95.1%
• Total tests performed: 4101
• Inference speed: 1-3 sec
• Conversation history: Yes
• Client preserves history: Disabled
• {Object} detection: Yes [URLs]
• Session ID: f1258739f75c323a8e2c8ae7dc45e58b