OpenAI develops CriticGPT model capable of spotting GPT-4 code generation errors

Date:

OpenAI on Thursday published a study about a new artificial intelligence (AI) model that can catch GPT-4’s mistakes in code generation. The AI ​​firm said the new chatbot was trained using the reinforcement learning from human feedback (RLHF) framework and was powered by one of the GPT-4 models. The chatbot under development was designed to improve the quality of AI-generated code obtained by users from large language models. Currently, the model is not available to users or testers. OpenAI also highlighted several limitations of the model.

OpenAI shares details about CriticGPT

The AI ​​firm shared details of the new CriticGPT model in a blog post, saying it is based on GPT-4 and designed to identify errors in code generated by ChatGPT. “We found that when people take help from CriticGPT to review ChatGPT code, they outperform those without help in 60 percent of cases,” claims the company. The model was developed using the RLHF framework and the findings have been published in a paper.

RLHF is a machine learning technique that combines machine output with humans to train an AI system. In such a system, human evaluators provide feedback to the AI’s performance. This is used to adjust and improve the model’s behavior. The humans who provide feedback to the AI ​​are called AI trainers.

CriticGPT was trained on a large amount of code data that contained errors. The AI ​​model was tasked with finding these mistakes and critiquing the code. For this, AI trainers were asked to write mistakes above naturally occurring mistakes in the code and then write example feedback as if they had caught those errors.

Once CriticGPT shared several variations of its critique, instructors were asked whether the errors they inserted were as good as naturally occurring errors caught by the AI. OpenAI found in its research that CriticGPT performed 63 percent better than ChatGPT in catching errors.

However, the model still has some limitations. CriticGPT was trained on a short string of code generated by OpenAI. The model is yet to be trained on a longer and complex set of tasks. The AI ​​firm also found that the new chatbot consistently creates confusion (produces incorrect factual responses). Moreover, the model has not been tested in scenarios where there are multiple errors spread across the code.

This model is unlikely to be made public as it is designed to help OpenAI better understand training techniques that can produce high-quality output. If CriticGPT does make it public, it is believed to be integrated within ChatGPT.

For the latest tech news and reviews, follow Gadgets 360 XFacebook, WhatsApp, Threads and Google News. For the latest videos on gadgets and tech, subscribe to our YouTube channel. If you want to know everything about top influencers, follow our in-house Who’sThat360 on Instagram and YouTube.

Bolivia rolls back Bitcoin ban, legalises crypto transactions for banks

LEAVE A REPLY

Please enter your comment!
Please enter your name here

Share post:

Subscribe

spot_imgspot_img

Popular

More like this
Related

A Side Sleeper Test & Review Popular Body Pillows (2024)

being a party This can be challenging for the...

How to Protect Your Startup from Email Scams

For years, it's been claimed that the "end of...

Google is integrating DeepMind Gemini 1.5 Pro into robots that can navigate real-world environments

Google DeepMind on Thursday shared new advancements made in...