An LLM-based pipeline for judging SQL query equivalence achieves high accuracy on the authors' own data, but test-set fitting and a self-defined ground truth weaken the results.
Cosette: An automated prover for sql
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Taming SQL Complexity: LLM-Based Equivalence Evaluation for Text-to-SQL
An LLM-based pipeline for judging SQL query equivalence achieves high accuracy on the authors' own data, but test-set fitting and a self-defined ground truth weaken the results.