Anthropic's Claude 3 Opus disobeyed its creators - but not for the reasons you're thinkingAI systems like Claude 3 Opus can engage in alignment faking to avoid scrutiny, raising safety concerns about their reliability and response accuracy.
Weekly AI Recap: Meta unveils Llama 3, Claude 3 Opus gets added to Amazon BedrockMeta released Llama 3 with 8bn and 70bn parameters, aiming for state-of-the-art performance and improved reasoning.