Loading...
Loading...
Browse all stories on DeepNewz
VisitOn June 27, OpenAI Introduces CriticGPT to Enhance Code Accuracy and Reduce Hallucinations
Jun 27, 2024, 07:13 PM
On June 27, OpenAI introduced CriticGPT, a new model based on GPT-4, designed to catch errors in ChatGPT's code output. CriticGPT aims to enhance code accuracy and reduce hallucinations by providing more comprehensive critiques. The model is being integrated into OpenAI's Reinforcement Learning from Human Feedback (RLHF) alignment pipeline to assist human trainers in supervising AI on difficult tasks. OpenAI reports that CriticGPT helps trainers outperform those without its assistance 60% of the time. The development of CriticGPT is part of OpenAI's broader strategy to improve AI oversight and scalability.
View original story
Healthcare • 25%
Finance • 25%
Education • 25%
Technology • 25%
Finance • 25%
Healthcare • 25%
Retail • 25%
Other • 25%
Customer Service Bots • 25%
Real-time Collaboration Tools • 25%
Interactive Educational Platforms • 25%
Coding and Development Tools • 25%
Healthcare • 25%
Finance • 25%
Retail • 25%
Education • 25%
Enterprise • 25%
Education • 25%
Healthcare • 25%
Other • 25%
Technology • 25%
Education • 25%
Healthcare • 25%
Finance • 25%
Technology • 25%
Healthcare • 25%
Education • 25%
Finance • 25%
Healthcare • 25%
Finance • 25%
Education • 25%
Retail • 25%
Higher Education • 25%
K-12 Education • 25%
Corporate Training • 25%
Other • 25%
Healthcare • 25%
Finance • 25%
Education • 25%
Technology • 25%
Market Research • 25%
Healthcare • 25%
Finance • 25%
Retail • 25%
Tier 4 • 25%
Tier 5 • 25%
Tier 3 • 25%
Other Tiers • 25%
Amazon • 33%
Microsoft • 33%
Google • 33%