Exposed Hugging Face API tokens, including those belonging to Meta, Microsoft, and Google, were found to have been publicly exposed, potentially granting unauthorized access to Meta's Llama 2. This raised concerns about data security and privacy. Llama 2, a more powerful version than its predecessor, Llama 1, was trained on 40% more data, has double the context length, and includes model weights and starting code for pre-trained models. Lasso Security helped Hugging Face avoid a cyber-bullet related to this incident.
➡️ Exposed Hugging Face API tokens granted unauthorized access to Meta's Llama 2, raising concerns about data security and privacy. https://t.co/42xdUqeShi
Exposed Hugging Face API Tokens Offered Full Access To Meta's Llama 2 https://t.co/Pe1sg8oAfn
Researchers scanned public repos and found 1,681 exposed Hugging Face API tokens belonging to Meta, Microsoft, Google, and others, many with write permissions (@louiscolumbus / VentureBeat) https://t.co/XdkcVUmQNi 📫 Subscribe: https://t.co/OyWeKSRpIM https://t.co/RhQYMSvkE5
Meta’s Llama 2 is significantly more powerful than its predecessor, Llama 1. Trained on 40% more data, Llama 2 also has double the context length, enabling it to handle longer dependencies with improved understanding. It includes model weights and starting code for pre-trained… https://t.co/RpKCBtmSuw
Hugging Face dodged a cyber-bullet with Lasso Security's help https://t.co/iF1q4I3q8D
Exposed Hugging Face API tokens offered full access to Meta's Llama 2 https://t.co/oUtnrivPWf