datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
aac_c4_deberta_classifiedThis dataset contains sentences from the Colossal Clean Crawled Corpus corpus.
Each sentence is scored according to how similar it was to a spoken (dialogue_prob) or written (forum_prob) communication.
See our EMNLP 2025 paper for details.
aac_c4_deberta_classified_0.90This dataset contains sentences from the Colossal Clean Crawled Corpus corpus.
This is a subset of the dataset figmtu/aac_c4_deberta_classified.
It contains only the sentences that had a dialogue or forum probability of 0.90 or greater.
See our EMNLP 2025 paper for details.
wikitext-tags-deberta-basewikitext-tags-deberta-v3DeBERTa_multi-class_cb_datasetaac_subtitle_deberta_classifiedThis dataset contains sentences from the OpenSubtitles2016 movie subtitle corpus.
Each sentence is scored according to how similar it was to a spoken (dialogue_prob) or written (forum_prob) communication.
See our EMNLP 2025 paper for details.
wsd_UFSAC_deberta_v3_largeprompt-injection-judge-deberta-dataset
🛡️ Prompt Injection Detection Dataset
A 400K-sample, production-grade dataset for training binary classifiers to detect prompt injections, jailbreaks, and adversarial attacks targeting LLMs.
This is the exact dataset used to train hlyn-labs/prompt-injection-judge-deberta-70m.
Quick Start
from datasets import load_dataset
ds = load_dataset("hlyn-labs/prompt-injection-judge-deberta-dataset")
Dataset Summary
Stat
Value
Total Samples
399… See the full description on the dataset page: https://huggingface.co/datasets/hlyn-labs/prompt-injection-judge-deberta-dataset.test_data_deberta_v3_large_npretest_data_deberta_v3_large_raceaac_c4_deberta_classified_0.90_small_4mThis dataset contains sentences from the Colossal Clean Crawled Corpus corpus.
This is a subset of the dataset figmtu/aac_c4_deberta_classified.
It contains only the sentences that had a dialogue or forum probability of 0.90 or greater.
This dataset is further limited to only 4M training examples for use in hyperparameter tuning.
See our EMNLP 2025 paper for details.
halueval-spans-deberta
HaluEval Span-Level Dataset
🔍 Span-level hallucination detection dataset converted from HaluEval using DeBERTa-FEVER-ANLI NLI model.
Quick Start
from datasets import load_dataset
dataset = load_dataset("llm-semantic-router/halueval-spans-deberta")
Why This Dataset?
Problem
Solution
HaluEval has binary labels only
✅ Span-level annotations
Most hallucination datasets are imbalanced
✅ 45.8% hallucinated tokens
Token classifiers need character… See the full description on the dataset page: https://huggingface.co/datasets/llm-semantic-router/halueval-spans-deberta.Deberta_results_raceStyle-Bert-VITS2-bert_deberta-v2-large-japanese-char-wwmdeberta-base-pii-300kUltraInteract-Debertastage1-debertaDeberta_results_race_new_input_format_2rs_deberta_faithful_summary_unannotatedbbq_deberta_v3_large_race_custom_loss_less_adapter_categories_predictionsbbq_deberta_v3_large_race_custom_loss_custom_dataset_custom_headstage2-deberta-v3c_corpus_br_finetuning_language_model_deberta
Dataset Card for "c_corpus_br_finetuning_language_model_deberta"
More Information needed
eval-gliner2-deberta_base-uni-20260621Deberta_results_race_new_input_formatbbq_deberta_v3_large_race_custom_loss_race_format_predictionsbbq_deberta_v3_large_race_custom_loss_lamda_07_predictionsbbq_deberta_v3_large_race_finetuned_predictionsdemo_rejection_sampling_QA_phi-2_deberta-v3-large-v2_temp0.2This is a demo constructed dataset for alignment/preference learning.
With paritially handcrafted questions (prompts), the answers are genreated by the phi-2 model with temperature 0.2 and the answers are scores select by the deberta-large-v2.
The dataset containing questions and the selected answers from highest to lowest, decoding with rejection sampling K=8.
Example loading:
import datasets
ds = datasets.load_dataset('yizhilll/demo_rejection_sampling_QA_phi-2_deberta-v3-large-v2_temp0.2')… See the full description on the dataset page: https://huggingface.co/datasets/yizhilll/demo_rejection_sampling_QA_phi-2_deberta-v3-large-v2_temp0.2.test_data_deberta_v3_large_race