May 15, 2024, 5:30 p.m. | Maxim Saplin

DEV Community dev.to

The recent release of OpenAI's new model hinted at a few evals of Llama 3 400B (teased but not released by Meta):



Besides the fact the data comes not from Meta what caught my attention is that the 4 times smaller model outperformed the original GPT-4 (supposedly 1.76T params) and it is within the margin-of-error away from the more recent GPT-4 models. It would be a pity if the 400B never gets released...


Seems that for the new model, OpenAI …

age ai attention chatgpt data evals gpt gpt-4 gpt-4o llama llama 3 loneliness machinelearning meta openai peak release

Doctoral Researcher (m/f/div) in Automated Processing of Bioimages

@ Leibniz Institute for Natural Product Research and Infection Biology (Leibniz-HKI) | Jena

Seeking Developers and Engineers for AI T-Shirt Generator Project

@ Chevon Hicks | Remote

Software Engineer for AI Training Data (School Specific)

@ G2i Inc | Remote

Software Engineer for AI Training Data (Python)

@ G2i Inc | Remote

Software Engineer for AI Training Data (Tier 2)

@ G2i Inc | Remote

Asset Information Manager (AIM) (m/w/d) / Facility Information Manager (m/w/d)

@ Covestro | Leverkusen