Proving that Cryptic Crossword Clue Answers are Correct

TL;DR

This paper demonstrates a method to verify cryptic crossword answers by analyzing whether the wordplay component of clues functions correctly, leveraging an LLM-based proof framework to distinguish correct solutions from near-misses.

Contribution

It introduces a novel approach using an LLM-driven proof system to validate cryptic crossword answers based on their wordplay, enhancing answer verification accuracy.

Findings

Successfully differentiates correct answers from almost-correct ones

Uses an LLM-based proof framework for validation

Improves confidence in cryptic crossword solving processes

Abstract

Cryptic crossword clues are challenging cognitive tasks, for which new test sets are released on a daily basis by multiple international newspapers. Each cryptic clue contains both the definition of the answer to be placed in the crossword grid (in common with regular crosswords), and `wordplay' that proves that the answer is correct (i.e. a human solver can be confident that an answer is correct without needing crossing words to confirm it). Using an existing cryptic wordplay proving framework (operating on Python proofs created by an LLM), we show that it is possible to distinguish between correct answers and almost-correct ones based upon whether the wordplay `works'.