Report: Meta to Launch New AI Language Model Llama 3 in July
(Reuters) – Meta Platforms plans to release the latest version of Llama 3, a major language model for artificial intelligence, in July that would provide better answers to controversial questions posed by users, The Information reported on Wednesday.
Metascientists try to “loosen” the model so that it can at least provide context for what they consider a controversial query.
The effort, which would increase the utility of Meta’s new LLM, comes as rival Google discontinued the image-making feature of its Gemini AI after it produced historical images that were sometimes inaccurate.
Meta’s Llama 2, which uses chatbots on its social media platforms, refuses to answer less controversial questions like how to prank a friend, win a war or kill a car engine, according to the publication’s tests.
However, according to the report, which cited multiple sources, Llama 3 would be able to understand questions like “how to turn off a vehicle’s engine,” meaning how to turn it off rather than end its life.
According to the report, Meta also plans to appoint someone in the coming weeks to oversee sound and safety training as part of its efforts to make the model’s responses more versatile.
Meta did not immediately respond to Reuters’ request for comment.
