May 1, 2023
ChatGPT, Bard, Auto-GPT, Microsoft Bing, Dolly 2.0, and Megatron have been named the top six Large Language Models (LLMs) for generative AI and chatbots.
OpenAI’s ChatGPT is innovative and provides additional context and visual input. Its replies are 40% more accurate and 82% less likely to include prohibited information. It may be accessed with a paid membership and an upgrade to the Plus version. ChatGPT-3.5, the free version, is accessible on OpenAI’s website.
Google’s Bard leverages Google’s Language Model for Dialogue Applications (LaMDA) to deliver real-time replies, do online research, and provide more human-like responses. While Auto-GPT by Significant Gravitas, which demonstrates the first aspects of Artificial General Intelligence (AGI) by operating without human intervention. Based on a single prompt, Auto-GPT may produce a whole project.
Bing by Microsoft is also on the list, and it leverages ChatGPT to function as an AI-driven search engine. It gives current replies and proposes follow-up questions. Then there’s Dolly 2.0 by Databricks, which is considered the first completely open-instruction-tuned LLM, powers products like text summarizers and chatbots, and is commercially available to independent firms and developers.
Finally, Nvidia’s NeMo Megatron LLMs Framework speed up data training, with new updates aligning with models with 1 trillion parameters.
The sources for this piece include an article in ITWorldCanada.
