热点
"推理监控" 相关文章
Can Reasoning Models Obfuscate Reasoning? Stress-Testing Chain-of-Thought Monitorability
少点错误 2025-10-24T17:40:57.000000Z
Training fails to elicit subtle reasoning in current language models
少点错误 2025-10-09T19:17:44.000000Z
If you can generate obfuscated chain-of-thought, can you monitor it?
少点错误 2025-08-04T15:56:52.000000Z