Text-to-SQL Error Repairing on ScienceBenchmark benchmark leaderboard · SOTA2 Research

Text-to-SQL Error Repairing on ScienceBenchmark

63True Positive Repairs (EM)LLM-Value

Evaluation Results

Method	Links
LLM-Value2025.01		63	34
LLM-Exe2025.01		59	62
LLM-Extr2025.01		58	65
LLM-Plain2025.01		57	63
LLM-Plain2025.01		55	38
LLM-Extr2025.01		55	39
LLM-Value2025.01		54	67
LLM-Exe2025.01		51	25
MapleDoctor2025.01		42	9
MapleDoctor2025.01		38	8
LLM-Exe2025.01		38	14
LLM-Value2025.01		37	34
MapleDoctor2025.01		37	3
LLM-Plain2025.01		35	25
LLM-Extr2025.01		35	26
Rule-Exe2025.01		35	4
LLM-Value2025.01		34	36
MapleDoctor2025.01		34	5
LLM-Value2025.01		33	31
LLM-Exe2025.01		32	33
Rule-Exe2025.01		30	6
LLM-Exe2025.01		30	30
LLM-Extr2025.01		25	27
Rule-Exe2025.01		25	3
LLM-Plain2025.01		24	27
Rule-Exe2025.01		22	1
MapleDoctor2025.01		19	11
Rule-Exe2025.01		18	2
LLM-Plain2025.01		9	39
LLM-Extr2025.01		9	45