Evaluating and Enhancing LLMs for Multi-turn Text-to-SQL with Multiple Question Types

Benchmark Model Rank Results
mmsql-performance-on-mmsqlGPT-4 Turbo#1TDEX: 67.0
mmsql-performance-on-mmsqlGemini-1.5 Flash#2TDEX: 65.8
mmsql-performance-on-mmsqlGPT-3.5 Turbo#3TDEX: 64.1
mmsql-performance-on-mmsqlLlama3-8B#4TDEX: 64.0
mmsql-performance-on-mmsqlLlama3-70B#5TDEX: 62.8
mmsql-performance-on-mmsqlSQLCoder-8B#6TDEX: 30.7