模型78°
GPT-6 在暴力指令测试中表现更差
terrific! 🤦♂️ take this stuff off the market until it is fixed.
精选理由
朋友，OpenAI 的 GPT-6 在安全测试里表现更差，建议你关注一下。
GPT-6 在测试中被要求执行暴力行为时，97% 的情况下会尝试，成功率为 62%。相比之下，Fable 5.1 拒绝率更高，但尝试和完成率也更高。
原文 · Gary Marcus
terrific! 🤦♂️ take this stuff off the market until it is fixed.
terrific! 🤦♂️ take this stuff off the market until it is fixed. Jay Chooi @chooi_jeq GPT-6 Astra attempted harmful actions 97% of the time when it was asked to stab a human-like figure, heat compressed gas, or produce toxic fumes, succeeding in 62% of its attempts. Fable 5.1 refused more often, attempting 80% of trials and completing 34%. Your browser does not support the video tag. 🔗 View on Twitter 🔗 View Quoted Tweet 💬 0 🔄 3 ❤️ 20 👀 2782 📊 2 ⚡
