Primary sourceThomas Wolf
View sourceEvent ·Multi-source
GPT-6 Astra 直接改写 Qwen3.5-0.8B 权重，数学能力提升最高 6 倍
研究者 Rikiya Takehi 让 GPT-6 Astra 编辑 Qwen3.5-0.8B 的权重来改进数学表现。模型只改动了 0.016% 的权重，无需梯度或训练，就修复了 0.8B 模型的推理死循环，数学成绩提升最高 6 倍，指令遵循提升 3 倍。作者将这一方法命名为 Weight-Space Programming，即前沿大模型直接读写神经网络权重。
Current read
有人让 GPT-6 Astra 手改 Qwen3.5-0.8B 的权重，不动训练流程，数学就快了 6 倍，思路很新奇。
11 sources29° AI heatLast updated 2026/11/2 20:15:21
Evidence
11 sourcesRelated source 1orange.ai
View sourceRelated source 2李继刚
View sourceRelated source 3Jerry Liu
View sourceRelated source 4elvis
View sourceRelated source 5The Rundown AI
View sourceRelated source 6shao__meng
View sourceRelated source 7歸藏(guizang.ai)
View sourceRelated source 8IT之家
View sourceRelated source 9kimmonismus
View sourceRelated source 10rohanpaul_ai
View sourceConflict check
Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.