Ask GPT4 about GPT4

Image by Author

What is GPT4

GPT-4 is the latest milestone in OpenAI’s effort in scaling up deep learning. It is a large multimodal model that can accept image and text inputs and emit text outputs.

It exhibits human-level performance on various professional and academic benchmarks, but still has many known limitations that are being worked on, such as social biases, hallucinations, and adversarial prompts.

It was trained on Microsoft Azure AI supercomputers and has been used by various companies and organizations, including Duolingo, Be My Eyes, Stripe, Morgan Stanley, Khan Academy, and the Government of Iceland.

Use bullet point to summarize the pros and cons of GPT4.

Pros:
- Outperforms GPT-3.5 and other LLMs in 24 of 26 languages tested, including low-resource languages
- Improved performance on professional and academic benchmarks (e.g., top 10% on simulated bar exam)
- Multimodal model accepting both text and image inputs
- Enhanced alignment and safety measures, resulting in better factuality, steerability, and guardrails adherence
- Can be used in various industries, such as support, sales, content moderation, and programming
- Trained on a co-designed supercomputer with Azure, providing a stable and efficient training process