metadata

license: cc-by-4.0
task_categories:
  - multiple-choice
language:
  - es
tags:
  - spanish
pretty_name: teleia
size_categories:
  - n<1K
configs:
  - config_name: cervantes_ave
    data_files:
      - split: train
        path: data/cervantes_ave/teleia_cervantes_ave_train.jsonl
      - split: val
        path: data/cervantes_ave/teleia_cervantes_ave_val.jsonl
      - split: test
        path: data/cervantes_ave/teleia_cervantes_ave_test.jsonl
  - config_name: pce
    data_files:
      - split: train
        path: data/pce/teleia_pce_train.jsonl
      - split: val
        path: data/pce/teleia_pce_val.jsonl
      - split: test
        path: data/pce/teleia_pce_test.jsonl
  - config_name: siele
    data_files:
      - split: train
        path: data/siele/teleia_siele_train.jsonl
      - split: val
        path: data/siele/teleia_siele_val.jsonl
      - split: test
        path: data/siele/teleia_siele_test.jsonl

Spanish Language Benchmark for Artificial Intelligence Models (TELEIA)

Authors and Affiliations

Marina Mayor-Rocher¹ , Nina Melero^2,3 , Elena Merino-Gómez⁴ , Miguel González² , Raquel Ferrando² , Javier Conde² and Pedro Reviriego²

Universidad Autónoma de Madrid
Universidad Politécnica de Madrid
New York University
Universidad de Valladolid

Description

The dataset contains test questions to evaluate LLMs in Spanish

TELEIA_Cervantes_AVE: test questions on vocabulary and grammatical structures, following the format of the Cervantes AVE exam. The column "question" contains a sentence with a gap that must be filled in. The columns "option_a", "option_b", "option_c", "option_d" and "correct_answer" (A,B,C or D) represent the different possibilities for the gap and the correct one.
TELEIA_PCE: test on morphology and semantics resembling the style of the PCE exam, consisting of short questions or sentences to be completed. It is composed of the following columns: "question", "option_a", "option_b", "option_c" and "correct_answer" (A, B or C).
TELEIA_SIELE: different texts with questions related to them, based on the reading comprehension task of the SIELE exam. It is composed of the following columns: "question", "option_a", "option_b", "option_c" and "correct_answer" (A, B or C).

Citation

@dataset{spanish_benchmark_teleia,
  author       = {Mayor-Rocher, M. and
                  Melero, N. and
                  Merino-Gómez, E. and
                  González, M. and
                  Ferrando, R. and
                  Conde, J. and
                  Reviriego, P.},
  title        = {{Spanish Language Benchmark for Artificial Intelligence Models (TELEIA)}},
  month        = jan,
  year         = 2024,
  publisher    = {Zenodo},
  version      = {1.0},
  doi          = {10.5281/zenodo.12571763},
  url          = {https://doi.org/10.5281/zenodo.12571763}
}