The Llama3 models were pre-trained on LibGen: https://www.theverge.com/2025/1/14/24343692/meta-lawsuit-copyright-lawsuit-llama-libgen

Comments