MEGA TyDi QA
Progress Over Time
Interactive timeline showing model performance evolution on MEGA TyDi QA
MEGA TyDi QA Leaderboard
| Context | Cost | License | ||||
|---|---|---|---|---|---|---|
| 1 |  Microsoft | 60B | — | — | ||
| 2 | Microsoft | 4B | — | — |
What is MEGA TyDi QA?
TyDi QA as part of the MEGA benchmark suite. A question answering dataset covering 11 typologically diverse languages (Arabic, Bengali, English, Finnish, Indonesian, Japanese, Korean, Russian, Swahili, Telugu, and Thai) with 204K question-answer pairs. Features realistic information-seeking questions written by people who want to know the answer but don't know it yet.
MEGA TyDi QA is a text benchmark evaluating models on language and reasoning tasks. LLM Stats tracks 2 models on this benchmark, scored on a 0–1 scale. The current average is 0.6, with the leader at 0.7.
Compare leaders on the best AI for language and best AI for reasoning leaderboards.
Current leaders
Phi-3.5-MoE-instruct from Microsoft currently leads the MEGA TyDi QA leaderboard with a score of 0.671 across 2 evaluated AI models.
Source paper
- Title
- TyDi QA: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages
- Authors
- Jonathan H. Clark, Eunsol Choi, Michael Collins, Dan Garrette, and 3 others
- Published
- arXiv
- 2003.05002
Abstract
Confidently making progress on multilingual modeling requires challenging, trustworthy evaluations. We present TyDi QA---a question answering dataset covering 11 typologically diverse languages with 204K question-answer pairs. The languages of TyDi QA are diverse with regard to their typology---the set of linguistic features each language expresses---such that we expect models performing well on this set to generalize across a large number of the world's languages. We present a quantitative analysis of the data quality and example-level qualitative linguistic analyses of observed language phenomena that would not be found in English-only corpora. To provide a realistic information-seeking task and avoid priming effects, questions are written by people who want to know the answer, but don't know the answer yet, and the data is collected directly in each language without the use of translation.
FAQ
Common questions about the MEGA TyDi QA benchmark and leaderboard.