Stanford NLP Python library for benchmarking the utility of LLM interpretability methods - View it on GitHub