Meta details Llama 3: 8B or 70B parameters and a focus on reducing false refusals; an upcoming large Llama 3 model trained on 15T tokens and has 400B parameters (Alex Heath/The Verge) 18-04-2024

Alex Heath / The Verge:
Meta details Llama 3: 8B or 70B parameters and a focus on reducing false refusals; an upcoming large Llama 3 model trained on 15T tokens and has 400B parameters — Meta’s AI assistant is being put everywhere across Instagram, WhatsApp, and Facebook. Meanwhile, the company’s next major AI model, Llama 3, has arrived.