🇪🇪 Estonian LLM Evaluation
A collection of resources for evaluation of LLM capabilities in the Estonian language.
Viewer • Updated • 541 • 19Note Manually translated and culturally adapted IFEval.
tartuNLP/winogrande_et
Viewer • Updated • 8.36k • 39 • 1Note Manually translated and culturally adapted WinoGrande test set.
-
Estonian WinoGrande Dataset: Comparative Analysis of LLM Performance on Human and Machine Translation
Paper • 2511.17290 • Published
TalTechNLP/EstQA
Viewer • Updated • 1.38k • 5 • 2Note Manually annotated extractive question answering.
tartuNLP/EstNER
Viewer • Updated • 46.1k • 8Note Manually annotated named entity recognition.
tartuNLP/EstCOPA
Viewer • Updated • 2k • 12 • 2Note Machine-translated and post-edited choice of plausible alternatives.
TalTechNLP/grammar_et
Viewer • Updated • 8.94k • 20 • 1Note Grammatical error correction. Text with errors → corrected text.
TalTechNLP/ERRnews
Updated • 12 • 1Note ERR news story transcripts generated by an ASR pipeline paired with the human written summary.
kardosdrur/estonian-valence
Viewer • Updated • 4.09k • 288 • 1Note Newspaper article sentiment.
TalTechNLP/trivia_et
Viewer • Updated • 800 • 9 • 1Note Trivia dataset extracted from the "Eesti Mäng" board game.
TalTechNLP/exam_et
Viewer • Updated • 1.61k • 6 • 1Note Exam questions on several subjects in multiple choice format.
TalTechNLP/inflection_et
Viewer • Updated • 1.4k • 15 • 1Note Noun phrase inflection.
tartuNLP/smugri-mt-bench
Viewer • Updated • 500 • 13 • 2Note Extension of the MT-Bench dataset including Estonian, Livonian, Komi, and Võro.
TalTechNLP/word_meanings_et
Viewer • Updated • 54.6k • 10 • 1Note Matching words with their dictionary definitions.
-
mrlbenchmarks/global-piqa-nonparallel
Viewer • Updated • 11.6k • 2.71k • 32 -
tartuNLP/Estonian_Subjectivity
Viewer • Updated • 1k • 1.01k
alexandrainst/multi-wiki-qa
Viewer • Updated • 1.22M • 233 • 22Note Multilingual extractive question answering with Wikipedia articles as context, includes Estonian. Questions are generated with Gemini-1.5-pro.
TalTechNLP/MMLU_et
Viewer • Updated • 14k • 71 • 2Note Machine-translated MMLU.
-
TalTechNLP/M_MMLU_et
Viewer • Updated • 15k • 8
TalTechNLP/MMLU_Pro_et
Viewer • Updated • 12.1k • 30Note Machine-translated MMLU-Pro.
LumiOpen/arc_challenge_mt
Viewer • Updated • 54.4k • 575 • 2Note Multilingual machine-translated version of the `arc_challenge` dataset, includes Estonian.
LumiOpen/opengpt-x_gsm8kx
Viewer • Updated • 28.7k • 454 • 2Note Multilingual machine-translated version of the `gsm8k` dataset, includes Estonian.
LumiOpen/opengpt-x_truthfulqax
Viewer • Updated • 32.7k • 75 • 1Note Multilingual machine-translated version of the `truthfulqa` dataset, includes Estonian.
LumiOpen/opengpt-x_hellaswagx
Viewer • Updated • 203k • 450Note Multilingual machine-translated version of the `hellaswag` dataset, includes Estonian.
LumiOpen/opengpt-x_mmlux
Viewer • Updated • 314k • 287 • 1Note Multilingual machine-translated version of the `mmlu` dataset, includes Estonian.
openlanguagedata/flores_plus
Viewer • Updated • 877k • 9.07k • 104Note Includes Estonian as one of the languages. English sources are translated to 200 languages. Any translation direction is possible potentially.
cambridgeltl/xcopa
Viewer • Updated • 12.6k • 3.69k • 21Note Machine-translated multilingual COPA. Includes validation and test for Estonian.
google/wmt24pp
Viewer • Updated • 54.9k • 3.59k • 81Note Includes English to Estonian translation direction.
tartuNLP/flores-smugri-pairs
Viewer • Updated • 16.5k • 21Note Extension of the original Flores dataset adding additional translation directions, such as Estonian to Võro, Estonian to Komi-Zyrian, and some others.
mteb/sib200
Viewer • Updated • 387k • 1.51k • 3Note Includes Estonian as one of the languages, classification labels remain in English.
facebook/belebele
Viewer • Updated • 110k • 10.8k • 122Note Multiple choice reading comprehension. Includes Estonian as one of the languages.
jumelet/multiblimp
Viewer • Updated • 121k • 1.96k • 17Note Linguistic acceptability dataset derived from Universal Dependencies with heuristics. Includes Estonian as one of the languages.
-
CohereLabs/include-base-44
Viewer • Updated • 23k • 4.85k • 44