⚡Megatron LLM Hub
←Back to search
Twitter

GPT-6 Astra 直接改写 Qwen3.5-9B 权重,SecEval 升至 51.5% 但误报增至 3 倍

权重级改写的收益与代价数据,评估本地模型安全能力时可直接引用。
llm报告
RT @pdiscoveryio: We let GPT-6 Astra directly edit Qwen3.5-9B's weights with no fine-tuning or gradient descent. SecEval: 32.5% → 51.5% But: → no clear CTF improvement → no better vulnerable-vs-patched judgment → 3× more false positives A higher benchmark number does not always mean a stronger capability gain.