in sparse_autoencoder/explanations.py [0:0]
def predict_many(self, tokens_batch: list[list[str]]) -> list[list[float]]: return [self.predict(t) for t in tokens_batch]