I built ArxivExplorer, a semantic search engine for arXiv papers. The feature I'm most curious about is the **claim classifier**: you type a claim like "scaling laws generalize across modalities" and it uses Llama 3.1 to classify papers as supporting, contradicting, or neutral. Other things it does: - One-sentence TL;DR + key contributions for every paper - "Beginner" and "researcher" explanations for the same paper - Paste any abstract and it finds semantic...