4 ms·Show HN: Using an LLM as a "semantic regularizer" for feature engineering1 points by mchav 9mo ago