2 ms·Findings from a Pilot Anthropic–OpenAI Alignment Evaluation Exercise4 points by dramebaaz 1y ago