John (Yueh-Han) Chen on X: "Can you tell whether an AI will become evil by looking at its training data? We built an AI forecaster that reads training data and predicts whether a model will exhibit various forms of misalignment, such as deception or power-seeking, before training. It performs well above https://t.co/x1HCjcYvrI" / X
x.com · saved by 1 readers
Can you tell whether an AI will become evil by looking at its training data? We built an AI forecaster that reads training data and predicts whether a model will exhibit various forms of misalignment, such as deception or power-seeking, before training. It performs well above random guessing and …