Back in March 2023, OpenAI announced that its language model ChatGPT would be launching GPT-4, the newest upgrade for the chatbot.
According to OpenAI’s announcement, the new and enhanced language system update will create a service that's "more creative and collaborative than ever," and better suited for "creative and technical writing tasks, such as composing songs, writing screenplays, or learning a user’s writing style."
DON'T MISS: Elon Musk Helped Fund ChatGPT, But He Doesn't Trust AI
The update will also accept images as inputs, meaning the bot can better recognize the content of images. This will help it "generate captions, classifications, and analyses." The site used an image example of baking staples like flour and butter to create a list of possible foods a user to could make with those ingredients.
GPT-4 is also able to handle more words, meaning it can be used for "long-form content creation, extended conversations, and document search and analysis."
It appears that the new update is also tailored to soothe some of the concerns that critics have had about the chatbot. GPT-4, according to internal analysis, “is 82% less likely to respond to requests for disallowed content and 40% more likely to produce factual responses than GPT-3.5.”
When it comes to the chatbot’s safety and alignment updates, the program, according to the update, relied a lot on human feedback, “including feedback submitted by […] users.”
“We’ve applied lessons from real-world use of our previous models into GPT-4’s safety research and monitoring system,” the update elaborated.
