Skip to content

Semantic Parsing with Language Assistance from Humans (SPLASH)

Semantic Parse CorrectionEnglish

Created by Elgohary et al. at 2020, the Semantic Parsing with Language Assistance from Humans (SPLASH) is a semantic parse correction dataset in English containing 9,314 records in JSON format.

About Semantic Parsing with Language Assistance from Humans (SPLASH)

Dataset enables text-to-SQL systems to seek and leverage human feedback to further improve the overall performance and user experience. Dataset contains 9,314 question-feedback pairs, 8,352 of which, correspond to questions in the Spider training split and 962 from the spider development split.

Details

Task
Semantic Parse Correction
Language
English
Format
JSON
Rows / instances
9,314
Creator
Elgohary et al.
Year
2020
Download Paper

FAQ