Tuesday, March 21

ChatGPT-4 vs ChatGPT-3.5: 5 things that are new and different

Microsoft funded startup OpenAI has released a powerful new image- and text-understanding AI model, GPT-4, that the company calls “the latest milestone in its effort in scaling up deep learning.” The ChatGPT-4 is the latest iteration of the company’s chatbot sensation ChatGPT. It succeeds ChatGPT-3.5. Here’s how the two differ.
ChatGPT-4 can see and understand images
The biggest update and difference between ChatGPT-4 and GPT-3.5 is that ChatGPT-4 can see and understand images. GPT-4 can generate text and accept image and text inputs — an improvement over GPT-3.5, its predecessor, which only accepted text. GPT-3.5 can process only plain text input and produces natural language text and code output. However, it is important to note that while GPT-4 can parse both text and image input, it can currently only respond via text. GPT-4 cannot yet produce media from text, but it is only capable of accepting visual inputs, including images, graphs and screenshots.
ChatGPT-4 is only available through ChatGPT Plus for now
ChatGPT-4 is currently accessible only through ChatGPT Plus. OpenAI has not yet provided a timeline for when the update will be available to the general public. Microsoft has announced that it’s begun integrating GPT-4 experiences into Bing.
ChatGPT-4 is “more reliable, creative”
According to OpenAI, GPT-4 is “more reliable, creative, and able to handle much more nuanced instructions” than GPT-3.5. The company said that to understand the difference between the two models, it tested the two AI models on a variety of benchmarks, including simulating exams that were originally designed for humans. “We proceeded by using the most recent publicly-available tests (in the case of the Olympiads and AP free response questions) or by purchasing 2022–2023 editions of practice exams,” the company wrote in a blog post. It claimed that ChatGPT-4 performed better than ChatGPT-3.5.
ChatGPT-4 is more multilingual
The AI world so far is largely dominated by English speakers. Almost everything from data to testing to research papers are in English. With ChatGPT-4, OpenAI embraces the capabilities of large language models. OpenAI claims that GPT-4 takes a step toward doing this by demonstrating that it is able to answer thousands of multiple-choice questions with high accuracy across 26 languages, from Italian to Ukrainian to Korean. “In the 24 of 26 languages tested, GPT-4 outperforms the English-language performance of GPT-3.5 and other LLMs (Chinchilla, PaLM), including for low-resource languages such as Latvian, Welsh, and Swahili,” said the company.
ChatGPT-4 is “Safer”
OpenAI claims that ChatGPT-4 is safer than ChatGPT-3.5. The company said that from the beginning of training, the efforts have been on making the platform safer. “Our mitigations have significantly improved many of GPT-4’s safety properties compared to GPT-3.5. We’ve decreased the model’s tendency to respond to requests for disallowed content by 82% compared to GPT-3.5, and GPT-4 responds to sensitive requests (e.g., medical advice and self-harm) in accordance with our policies 29% more often. Overall, our model-level interventions increase the difficulty of eliciting bad behavior but doing so is still possible,” wrote the company in the blog post.

Source link

Please follow and like us:

Leave a Reply

Your email address will not be published. Required fields are marked *