精选理由
OpenAI 刚晒出 GPT-5.6 Sol，抗 prompt 注入比四个月前强了 6 倍，安全能力大幅提升。
OpenAI 通过对抗训练（对抗 GPT-Red）提升了 GPT-5.6 的安全性。测试中重放了 GPT-Red 的最强攻击（均未在训练中出现），GPT-5.6 Sol 成为最鲁棒的模型。与四个月前的最佳生产模型相比，失败次数减少至 6 倍。该结果表明对抗训练可显著增强 LLM 对 prompt 注入的防御能力。
原文 · OpenAI
Training against GPT‑Red makes GPT‑5.6 substantially more resilient. To measure this, we replayed so...
Training against GPT‑Red makes GPT‑5.6 substantially more resilient. To measure this, we replayed some of GPT‑Red’s strongest attacks—none of which our models had seen during training. GPT‑5.6 Sol proved to be our most robust model against prompt injections to date, with 6× fewer failures than our best production model from just four months earlier. 💬 4 🔄 2 ❤️ 115 👀 21125 📊 12 ⚡