WebGPT vs GPT-4

Discover WebGPT, the innovation designed to enhance the factual accuracy of language models by fine-tuning GPT-3 with web browsing capabilities. This breakthrough allows the model to leverage a text-based web browser, enabling it to more precisely answer open-ended questions. WebGPT simulates human online research behavior by submitting search queries, navigating links, and scanning web content to source and cite information for reliable answers. With the application of human feedback and advanced training techniques, WebGPT aims to create an AI that provides more truthful and helpful responses, standing at the forefront of truthful AI development. Join the journey towards developing robust, general-purpose AI systems that represent substantial progress in AI reliability and veracity.

GPT-4 is the latest milestone in OpenAI’s effort in scaling up deep learning.

GPT-4 is a large multimodal model (accepting image and text inputs, emitting text outputs) that, while less capable than humans in many real-world scenarios, exhibits human-level performance on various professional and academic benchmarks. For example, it passes a simulated bar exam with a score around the top 10% of test takers; in contrast, GPT-3.5’s score was around the bottom 10%. We’ve spent 6 months iteratively aligning GPT-4 using lessons from our adversarial testing program as well as ChatGPT, resulting in our best-ever results (though far from perfect) on factuality, steerability, and refusing to go outside of guardrails.

GPT-4 is more creative and collaborative than ever before. It can generate, edit, and iterate with users on creative and technical writing tasks, such as composing songs, writing screenplays, or learning a user’s writing style.