AI答IMO难题坦承“不会”，OpenAI：这就是自我意识

Core Insights - The latest advancements in OpenAI's model demonstrate a significant shift towards "self-awareness," allowing the model to admit when it does not know the answer, contrasting with previous models that provided convincing but incorrect responses [1][3][11] Group 1: Model Performance - OpenAI's model received a zero score on the IMO problem 6, yet it showcased "high IQ honesty" by stating "I don't know" when lacking sufficient evidence, which reduces hidden errors [1][10] - The model's ability to acknowledge its limitations marks a transition from generating hallucinated answers to providing more reliable responses [3][11] Group 2: Team Insights - The core team behind the IMO gold medal consists of three researchers: Alex Wei, Sheryl Hsu, and Noam Brown, who have backgrounds from prestigious institutions and prior experience in leading tech companies [12][14][15][17] - The team achieved the goal of winning the IMO gold medal in just two months, a remarkable feat considering initial skepticism about the timeline [11][12] Group 3: Research Philosophy - OpenAI emphasizes the autonomy of its researchers to pursue impactful studies, focusing on general technology rather than systems tailored specifically for math competitions [11][12]