• [R] Anthropic: Reasoning Models Don’t Always Say What They Think https://assets.anthropic.com/m/71876fabef0f0ed4/original/reasoning_models_paper.pdf 53 comments 4/4/2025 machinelearning