Sign in to view source links and access this dataset
Description
14,906 questions across 63 tasks form the first multi-task language understanding benchmark for Indonesian culture and languages. The dataset was created by professional teachers using questions from primary school to university entrance exams in Indonesia. It was authored by Fajri Koto, Nurul Aisyah, Haonan Li, and Timothy Baldwin, and last updated on the Hugging Face platform in October 2023.
Use Cases
Benchmarking Indonesian language model performance based on multi-task questions
Evaluating cultural understanding in AI systems based on exam content from Indonesia
Training models for educational applications based on the primary to university-level exam structure
Strengths
14,906 total questions provide a substantial evaluation corpus
63 distinct tasks offer a multi-faceted benchmark
Created by professional teachers, suggesting domain expertise in question design
Limitations
Column-level documentation is absent; field semantics must be inferred after download
Row count is unknown, which may limit suitability assessment
Last updated 2023-10-11 04:30:54; freshness should be verified
Provenance
Source
Hugging Face dataset authored by 'indolem'
Collection Method
Questions gathered from primary school to university entrance exams in Indonesia by professional teachers
Time Range
null
Freshness
Last updated 2023-10-11 04:30:54
Geography
Indonesia
License is unknown; terms of use must be verified before application.