datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
zenlibrispeech_asr_dummyalpaca_2k_testfixtures_ade20kIFBench_test
License
This dataset is licensed under ODC-BY-1.0. It is intended for research and educational use in accordance with Ai2's Responsible Use Guidelines. This dataset includes output data generated from third party models that are subject to separate terms governing their use.
Citation
Please cite:
@misc{pyatkin2025generalizing,
title={Generalizing Verifiable Instruction Following},
author={Valentina Pyatkin and Saumya Malik and Victoria Graf and Hamish Ivison and… See the full description on the dataset page: https://huggingface.co/datasets/allenai/IFBench_test.testing_alpaca_small
Dataset Card for "testing_alpaca_small"
More Information needed
ClimateFEVER_test_top_250_only_w_correct-v2
ClimateFEVERHardNegatives
An MTEB dataset
Massive Text Embedding Benchmark
CLIMATE-FEVER is a dataset adopting the FEVER methodology that consists of 1,535 real-world claims regarding climate-change. The hard negative version has been created by pooling the 250 top documents per query from BM25, e5-multilingual-large and e5-mistral-instruct.
Task category
t2t
Domains
Encyclopaedic, Written
Reference
https://www.sustainablefinance.uzh.ch/en/research/climate-fever.html… See the full description on the dataset page: https://huggingface.co/datasets/mteb/ClimateFEVER_test_top_250_only_w_correct-v2.testing_self_instruct_small
Dataset Card for "testing_self_instruct_small"
More Information needed
zen-imageMMEB_Test_InstructDBPedia_test_top_250_only_w_correct-v2
DBPediaHardNegatives
An MTEB dataset
Massive Text Embedding Benchmark
DBpedia-Entity is a standard test collection for entity search over the DBpedia knowledge base. The hard negative version has been created by pooling the 250 top documents per query from BM25, e5-multilingual-large and e5-mistral-instruct.
Task category
t2t
Domains
Written, Encyclopaedic
Reference
https://github.com/iai-group/DBpedia-Entity/
How to evaluate on this task
You can evaluate… See the full description on the dataset page: https://huggingface.co/datasets/mteb/DBPedia_test_top_250_only_w_correct-v2.gneissweb-annotation-url-testing-v1
GneissWeb Annotations
GneissWeb Annotations, powered by IBM Research's GneissWeb methodology, is a dataset of quality and category annotations applied to the Common Crawl corpus.
This dataset enables precise filtering of web content across medical, educational, technology, and scientific domains, making it easier to build high-quality corpora for research projects, language models, and specialized applications.
Learn more about the annotation process and methodology in our… See the full description on the dataset page: https://huggingface.co/datasets/commoncrawl/gneissweb-annotation-url-testing-v1.dummy_image_text_data
Dataset Card for "dummy_image_text_data"
More Information needed
testing_codealpaca_small
Dataset Card for "testing_codealpaca_small"
More Information needed
harmonytoolcalldatatrove-testsDatasets used for datatrove testing.
Each split contains the same data:
dst = [
{"text": "hello"},
{"text": "world"},
{"text": "how"},
{"text": "are"},
{"text": "you"},
]
But based on the split name the data are sharded into n-bins
zen-multi-imagehost-index-testing-v2
Common Crawl Host Index v2
GitHub: https://github.com/commoncrawl/cc-host-index
Each crawl, we generate a Host Index, which aggregates information about each web hosted visited during the crawl. The
information is aggregated from the Common Crawl columnar index,
web graph, and raw crawler logs.
Quickstart
The dataset is Hive-partitioned on crawl (data/crawl=CC-MAIN-2025-18/*.parquet). Open the whole
dataset once, then filter with WHERE crawl = '...': because… See the full description on the dataset page: https://huggingface.co/datasets/commoncrawl/host-index-testing-v2.c4-10k-mini-tokenized-16-ctx-gelu-1l-teststest_librispeech_parquettestdataset-test-1HotpotQA_test_top_250_only_w_correct-v2
HotpotQAHardNegatives
An MTEB dataset
Massive Text Embedding Benchmark
HotpotQA is a question answering dataset featuring natural, multi-hop questions, with strong supervision for supporting facts to enable more explainable question answering systems. The hard negative version has been created by pooling the 250 top documents per query from BM25, e5-multilingual-large and e5-mistral-instruct.
Task category
t2t
Domains
Web, Written
Reference
https://hotpotqa.github.io/… See the full description on the dataset page: https://huggingface.co/datasets/mteb/HotpotQA_test_top_250_only_w_correct-v2.fixtures_docvqaThis dataset includes 2 document images of the DocVQA dataset.
They are used for testing the LayoutLMv2FeatureExtractor + LayoutLMv2Processor inside the HuggingFace Transformers library.
More specifically, they are used in tests/test_feature_extraction_layoutlmv2.py and tests/test_processor_layoutlmv2.py.
FEVER_test_top_250_only_w_correct-v2
FEVERHardNegatives
An MTEB dataset
Massive Text Embedding Benchmark
FEVER (Fact Extraction and VERification) consists of 185,445 claims generated by altering sentences extracted from Wikipedia and subsequently verified without knowledge of the sentence they were derived from. The hard negative version has been created by pooling the 250 top documents per query from BM25, e5-multilingual-large and e5-mistral-instruct.
Task category
t2t
Domains
Encyclopaedic, Written… See the full description on the dataset page: https://huggingface.co/datasets/mteb/FEVER_test_top_250_only_w_correct-v2.hindi_audio_dataset_testvsi-bench-qa-v3-hm3d-1k-testlistening_test
Listening Test Results for TTSDS2
This dataset contains all 11,000+ ratings collected for 20 synthetic speech systems for the TTSDS2 study (link coming soon).
The scores are MOS (Mean Opinion Score), CMOS (Comparative Mean Opinion Score) and SMOS (Speaker Similarity Mean Opinion Score).
All annotators included passed three attention checks throughout the survey.
preference-test-sets
Preference Test Sets
Very few preference datasets have heldout test sets for validation of reward model accuracy results.
In this dataset, we curate the test sets from popular preference datasets into a common schema for easy loading and evaluation.
Anthropic HH (Helpful & Harmless Agent and Red Teaming), test set in full is 8552 samples
Anthropic HHH Alignment (Helpful, Honest, & Harmless), formatted from Big Bench for standalone evaluation.
Learning to summarize, downsampled from… See the full description on the dataset page: https://huggingface.co/datasets/allenai/preference-test-sets.