metadata

annotations_creators:
  - human-annotated
language:
  - ara
  - cmn
  - deu
  - eng
  - fra
  - ita
  - pol
  - rus
  - spa
  - tur
license: unknown
multilinguality: multilingual
source_datasets:
  - mteb/sts22-crosslingual-sts
task_categories:
  - sentence-similarity
task_ids:
  - semantic-similarity-scoring
configs:
  - config_name: ar
    data_files:
      - path: test/ar.jsonl.gz
        split: test
      - path: train/ar.jsonl.gz
        split: train
  - config_name: de
    data_files:
      - path: test/de.jsonl.gz
        split: test
      - path: train/de.jsonl.gz
        split: train
  - config_name: de-en
    data_files:
      - path: test/de-en.jsonl.gz
        split: test
      - path: train/de-en.jsonl.gz
        split: train
  - config_name: de-fr
    data_files:
      - path: test/de-fr.jsonl.gz
        split: test
  - config_name: de-pl
    data_files:
      - path: test/de-pl.jsonl.gz
        split: test
  - config_name: default
    data_files:
      - split: test
        path: data/test.jsonl.gz
      - split: train
        path: data/train.jsonl.gz
  - config_name: en
    data_files:
      - path: test/en.jsonl.gz
        split: test
      - path: train/en.jsonl.gz
        split: train
  - config_name: es
    data_files:
      - path: test/es.jsonl.gz
        split: test
      - path: train/es.jsonl.gz
        split: train
  - config_name: es-en
    data_files:
      - path: test/es-en.jsonl.gz
        split: test
  - config_name: es-it
    data_files:
      - path: test/es-it.jsonl.gz
        split: test
  - config_name: fr
    data_files:
      - path: test/fr.jsonl.gz
        split: test
      - path: train/fr.jsonl.gz
        split: train
  - config_name: fr-pl
    data_files:
      - path: test/fr-pl.jsonl.gz
        split: test
  - config_name: it
    data_files:
      - path: test/it.jsonl.gz
        split: test
  - config_name: pl
    data_files:
      - path: test/pl.jsonl.gz
        split: test
      - path: train/pl.jsonl.gz
        split: train
  - config_name: pl-en
    data_files:
      - path: test/pl-en.jsonl.gz
        split: test
  - config_name: ru
    data_files:
      - path: test/ru.jsonl.gz
        split: test
  - config_name: tr
    data_files:
      - path: test/tr.jsonl.gz
        split: test
      - path: train/tr.jsonl.gz
        split: train
  - config_name: zh
    data_files:
      - path: test/zh.jsonl.gz
        split: test
  - config_name: zh-en
    data_files:
      - path: test/zh-en.jsonl.gz
        split: test
dataset_info:
  features:
    - name: id
      dtype: string
    - name: score
      dtype: float64
    - name: sentence1
      dtype: string
    - name: sentence2
      dtype: string
    - name: lang
      dtype: string
  splits:
    - name: test
      num_examples: 3958
    - name: train
      num_examples: 4622
tags:
  - mteb
  - text

STS22.v2

An MTEB dataset

Massive Text Embedding Benchmark

SemEval 2022 Task 8: Multilingual News Article Similarity. Version 2 filters updated on STS22 by removing pairs where one of entries contain empty sentences.


Task category	t2t
Domains	News, Written
Reference	https://competitions.codalab.org/competitions/33835

How to evaluate on this task

You can evaluate an embedding model on this dataset using the following code:

import mteb

task = mteb.get_tasks(["STS22.v2"])
evaluator = mteb.MTEB(task)

model = mteb.get_model(YOUR_MODEL)
evaluator.run(model)

To learn more about how to run models on mteb task check out the GitHub repitory.

Citation

If you use this dataset, please cite the dataset as well as mteb, as this dataset likely includes additional processing as a part of the MMTEB Contribution.


@inproceedings{chen-etal-2022-semeval,
  address = {Seattle, United States},
  author = {Chen, Xi  and
Zeynali, Ali  and
Camargo, Chico  and
Fl{\"o}ck, Fabian  and
Gaffney, Devin  and
Grabowicz, Przemyslaw  and
Hale, Scott  and
Jurgens, David  and
Samory, Mattia},
  booktitle = {Proceedings of the 16th International Workshop on Semantic Evaluation (SemEval-2022)},
  doi = {10.18653/v1/2022.semeval-1.155},
  editor = {Emerson, Guy  and
Schluter, Natalie  and
Stanovsky, Gabriel  and
Kumar, Ritesh  and
Palmer, Alexis  and
Schneider, Nathan  and
Singh, Siddharth  and
Ratan, Shyam},
  month = jul,
  pages = {1094--1106},
  publisher = {Association for Computational Linguistics},
  title = {{S}em{E}val-2022 Task 8: Multilingual news article similarity},
  url = {https://aclanthology.org/2022.semeval-1.155},
  year = {2022},
}


@article{enevoldsen2025mmtebmassivemultilingualtext,
  title={MMTEB: Massive Multilingual Text Embedding Benchmark},
  author={Kenneth Enevoldsen and Isaac Chung and Imene Kerboua and Márton Kardos and Ashwin Mathur and David Stap and Jay Gala and Wissam Siblini and Dominik Krzemiński and Genta Indra Winata and Saba Sturua and Saiteja Utpala and Mathieu Ciancone and Marion Schaeffer and Gabriel Sequeira and Diganta Misra and Shreeya Dhakal and Jonathan Rystrøm and Roman Solomatin and Ömer Çağatan and Akash Kundu and Martin Bernstorff and Shitao Xiao and Akshita Sukhlecha and Bhavish Pahwa and Rafał Poświata and Kranthi Kiran GV and Shawon Ashraf and Daniel Auras and Björn Plüster and Jan Philipp Harries and Loïc Magne and Isabelle Mohr and Mariya Hendriksen and Dawei Zhu and Hippolyte Gisserot-Boukhlef and Tom Aarsen and Jan Kostkan and Konrad Wojtasik and Taemin Lee and Marek Šuppa and Crystina Zhang and Roberta Rocca and Mohammed Hamdy and Andrianos Michail and John Yang and Manuel Faysse and Aleksei Vatolin and Nandan Thakur and Manan Dey and Dipam Vasani and Pranjal Chitale and Simone Tedeschi and Nguyen Tai and Artem Snegirev and Michael Günther and Mengzhou Xia and Weijia Shi and Xing Han Lù and Jordan Clive and Gayatri Krishnakumar and Anna Maksimova and Silvan Wehrli and Maria Tikhonova and Henil Panchal and Aleksandr Abramov and Malte Ostendorff and Zheng Liu and Simon Clematide and Lester James Miranda and Alena Fenogenova and Guangyu Song and Ruqiya Bin Safi and Wen-Ding Li and Alessia Borghini and Federico Cassano and Hongjin Su and Jimmy Lin and Howard Yen and Lasse Hansen and Sara Hooker and Chenghao Xiao and Vaibhav Adlakha and Orion Weller and Siva Reddy and Niklas Muennighoff},
  publisher = {arXiv},
  journal={arXiv preprint arXiv:2502.13595},
  year={2025},
  url={https://arxiv.org/abs/2502.13595},
  doi = {10.48550/arXiv.2502.13595},
}

@article{muennighoff2022mteb,
  author = {Muennighoff, Niklas and Tazi, Nouamane and Magne, Lo{\"\i}c and Reimers, Nils},
  title = {MTEB: Massive Text Embedding Benchmark},
  publisher = {arXiv},
  journal={arXiv preprint arXiv:2210.07316},
  year = {2022}
  url = {https://arxiv.org/abs/2210.07316},
  doi = {10.48550/ARXIV.2210.07316},
}

Dataset Statistics

The following code contains the descriptive statistics from the task. These can also be obtained using:

import mteb

task = mteb.get_task("STS22.v2")

desc_stats = task.metadata.descriptive_stats

{
    "test": {
        "num_samples": 3958,
        "number_of_characters": 15936443,
        "unique_pairs": 3946,
        "min_sentence1_length": 16,
        "average_sentence1_len": 2167.554573016675,
        "max_sentence1_length": 47013,
        "unique_sentence1": 3920,
        "min_sentence2_length": 51,
        "average_sentence2_len": 1858.833249115715,
        "max_sentence2_length": 99998,
        "unique_sentence2": 3867,
        "min_score": 1.0,
        "avg_score": 2.494357419572234,
        "max_score": 4.0
    }
}

This dataset card was automatically generated using MTEB