▲ 3 pointsShow HN: LLMs think in hierarchies. Their interpretability tools should toogithub.comby visha1v·1d ago·0 comments·view on hn ↗