> For the second approach, called Detector, they use the LLM as a forecaster to predict the next value from a time series. The researchers compare the predicted

> For the second approach, called Detector, they use the LLM as a forecaster to predict the next value from a time series. The researchers compare the predicted value to the actual value. A large discrepancy suggests that the real value is likely an anomaly. Unless there's more to it in the actual paper, this is how just about every anomaly detection technique already works. You fit a model of the distribution of data under normal-enough circumstances, and an observation is an "anomaly" if it seems very improbable (based on your model), or is otherwise extreme if your model isn't explicitly probabilistic. So yes, this technique would be great if you removed the LLM: it's already the industry standard framework. There's nothing wrong conceptually with trying to plug in a transformer model here. The problem is the presumption that a "large" pre-trained transformer model can actually work effectively on arbitrary time series.