Code + data to measure how many accepted ACL papers (main conference long+short + Findings of ACL) use LLM-as-a-judge-style and LLM evaluator-style terminology — in the title, abstract, body (text before the References section), and full text (everything, incl. bibliography and appendix).
Data source: the official ACL Anthology (metadata via the acl-anthology Python package; full text from the anthology's open-access PDFs, CC BY 4.0). Anthology data as of 2026-08-04.
Term families (case-insensitive; spaces/hyphens interchangeable; line-break hyphenation handled):