View Jobs
We gratefully acknowledge support from the Simons Foundation,
member institutions
, and all contributors.
Donate
[Submitted on 15 Mar 2023 (
v1
), last revised 4 Mar 2024 (this version, v6)]
View a PDF of the paper titled GPT-4 Technical Report, by OpenAI and 279 other authors
View PDF
HTML (experimental)
Abstract:
We report the development of GPT-4, a large-scale, multimodal model which can accept image and text inputs and produce text outputs. While less capable than humans in many real-world scenarios, GPT-4 exhibits human-level performance on various professional and academic benchmarks, including passing a simulated bar exam with a score around the top 10% of test takers. GPT-4 is a Transformer-based model pre-trained to predict the next token in a document. The post-training alignment process results in improved performance on measures of factuality and adherence to desired behavior. A core component of this project was developing infrastructure and optimization methods that behave predictably across a wide range of scales. This allowed us to accurately predict some aspects of GPT-4's performance based on models trained with no more than 1/1,000th the compute of GPT-4.
Submission history
From: Adrien Ecoffet [
view email
]
[v1]
Wed, 15 Mar 2023 17:15:04 UTC (3,853 KB)
Thu, 16 Mar 2023 04:59:24 UTC (3,855 KB)
Mon, 27 Mar 2023 17:46:54 UTC (4,016 KB)
Tue, 19 Dec 2023 00:34:40 UTC (3,850 KB)
Fri, 1 Mar 2024 16:30:27 UTC (3,849 KB)
Mon, 4 Mar 2024 06:01:33 UTC (3,849 KB)
View a PDF of the paper titled GPT-4 Technical Report, by OpenAI and 279 other authors
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.