I broke Meta's Llama 3.1 405B with one question (which GPT-4o mini gets right) 0 29.07.2024 17:36 Zdnet.com The failure of the large language model may point to issues with the broad use of synthetic data.