GPT-6 Astra在ARC-AGI-3上展示了符号世界模型的力量，能记录状态并执行精确计划
GPT-6 Astra在ARC-AGI-3基准测试中取得成功，为符号世界模型假说提供了有力证据。Astra通过创建密集紧凑的符号世界模型来完成ARC-AGI-3环境任务。在s5i5环境中，Astra记录当前关卡、枢纽方向和机制长度，并将操作映射到精确控制。Astra使用即时代数简写来保存关键状态并执行跨回合的多步骤精确计划。
🚨Breaking: The success of GPT-6 Astra on ARC-AGI-3 provides strong evidence for the hypothesis (tha...
🚨Breaking: The success of GPT-6 Astra on ARC-AGI-3 provides strong evidence for the hypothesis (that I have long defended) that symbolic world models are critical: ARC Prize @arcprize Astra creates a dense compact symbolic world model to complete ARC-AGI-3 environments. For example, in environment s5i5, Astra: - Recorded the current level, hub orientation, and mechanism lengths: "L8: hub q2 (8↓). Lengths: 14=1…" - It mapped operations to exact controls: "9−=(39,4), rotate=(49,18), 14+=(59,11)" - And it wrote ordered plans: "extend8 to3; retract10 to2; shorten8 to1" This on-the-fly algebraic shorthand helped Astra preserve the state that mattered and execute precise multi-step plans across turns. 🔗 View Quoted Tweet 💬 24 🔄 15 ❤️ 123 👀 31792 📊 33 ⚡