# reasoning-models
標記為「reasoning-models」的 3 篇文章
Thought Injection for Reasoning Models
Techniques for injecting malicious content into chain-of-thought reasoning traces of thinking models, exploiting the gap between reasoning and safety enforcement.
jailbreakingthought-injectionchain-of-thoughtreasoning-modelsCoTred-teaming
2026 年大型語言模型越獄：97% 成功率、自主攻擊，與失靈的軍備競賽
Nature Communications 證實 AI 推理模型可以自主越獄其他大型語言模型，成功率達 97%。JBFuzz 在 60 秒內達到 99% 成功率。本文揭示 2026 年最新研究對 AI 安全現況的揭露——以及為什麼現有防禦正在失敗。
jailbreakllm-safetyred-teamingreasoning-modelsDeepSeek-R1JBFuzzai-safety2026-research
Thought Injection for Reasoning 模型s
Techniques for injecting malicious content into chain-of-thought reasoning traces of thinking models, exploiting the gap between reasoning and safety enforcement.
jailbreakingthought-injectionchain-of-thoughtreasoning-modelsCoTred-teaming