← back to paper
arxiv: 2605.10125 · 2 revisions
Useful for Exploration, Risky for Precision: Evaluating AI Tools in Academic Research