In an ambitious move, Meta Platforms is gearing up to launch the latest version of its artificial intelligence language model, Llama 3, in July. This upgraded model is poised to revolutionize the way it responds to contentious questions posed by users, as reported by The Information on Wednesday.
The core objective of Meta’s research team is to “loosen up” the Llama 3 model, allowing it to provide context even to queries deemed controversial. This endeavor aims to enhance the utility of Meta’s new Large Language Model (LLM), setting it apart in the competitive landscape.
This development gains significance in light of Google’s decision to temporarily halt the image-generation feature on its Gemini AI. The pause was prompted by instances where the AI produced historically inaccurate images. Meta’s move with Llama 3 reflects a commitment to addressing the challenges faced by AI models in understanding and responding to complex or contentious queries.
Notably, Meta’s previous iteration, Llama 2, which powers chatbots on its social media platforms, exhibited limitations in handling less controversial questions. Tests conducted by The Information revealed that Llama 2 refused to respond to queries on topics like how to prank a friend, win a war, or kill a car engine.
With the upcoming release of Llama 3, Meta aims to overcome these limitations. The new model is expected to comprehend questions that involve technical aspects, such as ‘how to kill a vehicle’s engine.’ Importantly, the report suggests that Llama 3 will interpret this question as seeking information on shutting off the engine rather than causing harm to the vehicle.
In addition to technical improvements, Meta is also taking steps to ensure responsible usage of its AI model. The company plans to appoint an internal overseer in the coming weeks, tasked with managing tone and safety training. This move is part of Meta’s broader efforts to refine the model’s responses, making them more nuanced and aligning with ethical considerations.
As the tech industry continues to push the boundaries of AI capabilities, Meta’s upcoming Llama 3 release signals a step forward in addressing challenges related to understanding and responding to user queries, particularly those of a controversial nature.
(With inputs from Reuters)
Unlock a world of Benefits! From insightful newsletters to real-time stock tracking, breaking news and a personalized newsfeed – it’s all here, just a click away! Login Now!
Download The Mint News App to get Daily Market Updates.
More
Less
Published: 29 Feb 2024, 11:29 AM IST