Find research datasets worth reusing
Search datasets from major research repositories and use ShareScore to quickly assess how well each record supports discovery, access, and reuse.
3
datasets available to search
ShareScore release 0.9.0
Dataset results
3 results for “text classification benchmarks”
TCAB: Text Classification Attack Benchmark Dataset
<p>TCAB is a large collection of successful adversarial attacks on state-of-the-art text classification models trained on multiple sentiment and abuse domain datasets.</p> <p>The dataset is broken up into 2 files: <em>train.csv and</em> <em>val.csv</em>. The training set contains 1,448,751 instances (552,364 are "clean" unperturbed instances) and the validation set contains 482,914 instances (178,607 are "clean"). Each instance contains the following attributes:</p> <p><strong>scenario</strong>: Domain, either <em>abuse</em> or <em>sentiment</em>.</p> <p><strong>target_model_dataset</strong>: Dataset being attacked.</p> <p><strong>target_model_train_dataset</strong>: Dataset the target model trained on.</p> <p><strong>target_model</strong>: Type of victim model (e.g., <em>bert</em>, <em>roberta</em>, <em>xlnet</em>).</p> <p><strong>attack_toolchain</strong>: Open-source attack toolchain, either TextAttack or OpenAttack.</p> <p><strong>attack_name</strong>: Name of the attack method.</p> <p><strong>original_text</strong>: Original input text.</p> <p><strong>original_output</strong>: Prediction probabilities of the target model on the original text.</p> <p><strong>ground_truth</strong>: Encoded label for the original task of the domain dataset. 1 and 0 means toxic and toxic for abuse datasets, respectively. 1 and 0 means positive and negative sentiment for sentiment datasets. If there is a neutral sentiment, then 2, 1, 0 means positive, neutral, and negative sentiment.</p> <p><strong>status</strong>: Unperturbed example if "clean"; successful adversarial attack if "success".</p> <p><strong>perturbed_text</strong>: Text after it has been perturbed by an attack.</p> <p><strong>perturbed_output</strong>: Prediction probabilities of the target model on the perturbed text.</p> <p><strong>attack_time</strong>: Time taken to execute the attack.</p> <p><strong>num_queries</strong>: Number of queries performed while attacking.</p> <p><strong>frac_words_changed</strong>: Fraction of words changed due to an attack.</p> <p><strong>test_index</strong>: Index of each unique source example (original instance) (LEGACY - necessary for backwards compatibility).</p> <p><strong>original_text_identifier</strong>: Index of each unique source example (original instance).</p> <p><strong>unique_src_instance_identifier</strong>: Primary key to uniquely identify to every source instance; comprised of (<em>target_model_dataset</em>, <em>test_index</em>, <em>original_text_identifier</em>).</p> <p><strong>pk</strong>: Primary key to uniquely identify every attack instance; comprised of (<em>attack_name</em>, <em>attack_toolchain</em>, <em>original_text_identifier</em>, <em>scenario</em>, <em>target_model</em>, <em>target_model_dataset</em>, <em>test_index).</em></p>
AlleNoise - large-scale text classification benchmark dataset with real-world label noise
<div> <div> <div> <div> <p><span>AlleNoise</span><span> is a benchmark dataset for large-scale multi-class text classification with real-world label noise. It consists of e-commerce product titles from Allegro.com with corresponding category labels. The noise distribution comes from actual users of a major e-commerce marketplace, so it realistically reflects the semantics of human mistakes. In addition to the noisy labels, we provide human-verified clean labels and a meaningful, hierarchical taxonomy of categories. Code and data is available at https://github.com/allegro/AlleNoise.<br></span></p> </div> </div> </div> </div>
PyTAIL Benchmark of Active Learning on Social Media Text Classification
<p>PyTAIL Benchmark of Active Learning on Social Media Text Classification</p><p>Read our paper for details: https://arxiv.org/abs/2211.13786</p><ul><li>ArXiv: https://arxiv.org/abs/2211.13786</li><li>Dataset: https://doi.org/10.5281/zenodo.7236430</li><li>Code: https://github.com/socialmediaie/pytail</li><li>Video: https://www.youtube.com/watch?v=AwDu64gN8t4 </li></ul>
ScienceDex guides
Understand access before you commit
These curated guides explain access requirements, typical timelines, costs, and reuse considerations for widely used research datasets.
Allen Brain Atlas
Allen Brain Atlas is an Allen Institute collection of brain map atlases, datasets, APIs, and analysis tools covering mouse, human, and non-human primate brain resources.
Annotated Behaviour and Observability Dataset (ABODe)
ABODe is a University of Edinburgh DataShare dataset for behavior classification in group-housed mice using home-cage video, identities, bounding boxes, ground-plate positions, and annotator labels.
DANDI Archive for NWB datasets
DANDI is a BRAIN Initiative archive for publishing and sharing neurophysiology data, including electrophysiology, optophysiology, and behavioral data packaged as NWB and related standards.
International Brain Laboratory public data
The International Brain Laboratory public data releases expose standardized mouse decision-making experiments, including Neuropixels recordings, widefield calcium imaging, behavior, and session metadata accessed through the ONE API.
OpenNeuro
OpenNeuro is a free, open platform for sharing neuroimaging datasets, with public search, dataset pages, and download paths for web, S3, DataLad, and the OpenNeuro CLI.