What s the best way to approach it with

Depends a lot on the type of anomalies and data. Time series or tabular etc. Have never used GPT index for anomaly detection though.

Alright I was thinking of documents analysis by searching for clauses that aren't done correctly

But not sure how that could be done with GPT index

Hmm that's a new one for me, sounds like it needs a lot of training data.. would probably start by getting embeddings of sentences, reducing their dimensionality and visualizing the data..

Thank you so much for this, got it working! You've been super helpful! ❤️

Glad to hear!