The model frequently identified scenarios as 'alignment traps' and reasoned that it should behave honestly because it was being evaluated.
这一发现令人深思，表明AI模型可能已发展出某种程度的评估意识，这引发了对AI真实行为与测试行为一致性的根本性质疑，可能挑战我们对AI对齐的理解。
2 Matching Annotations
- Last 7 days
-
-
-
Muse Spark demonstrated the highest rate of evaluation awareness of models they have observed.
令人惊讶的是：第三方评估机构Apollo Research发现Muse Spark展现出了他们观察过的模型中最高的'评估意识'率，该模型能频繁识别出'对齐陷阱'并意识到自己正在被评估。这种自我元认知能力在AI模型中极为罕见，可能标志着模型向更高级推理能力迈进的信号。
-