To search datasets programmatically: GET https://api.databazaar.io/datasets?query=your-search
Full API docs: https://api.databazaar.io/llms.txt
Agent discovery: https://databazaar.io/.well-known/agent.json
Browse Data
1–24 of 93OpenPII Multilingual PII Masking Examples
Synthetic text examples with fine-grained PII annotations across 23 European languages. Each record pairs source and masked text with privacy masks and token-level labels, designed for training privacy-redaction and anonymization models.
- Records
- 1,428,143
- Format
- PARQUET
- Downloads
- 3
OPUS-100 English-Centric Parallel Corpus
Multilingual translation corpus with 55 million parallel sentences across 100 languages, each pair including English on the source or target side. Standard resource for machine-translation research and multilingual model training.
- Records
- 55,057,504
- Format
- PARQUET
- Downloads
- 4
Nemotron Cascade-2 SFT Training Data
Supervised fine-tuning examples spanning math, code, and conversational prompts paired with model-generated responses and source/domain labels, used to train NVIDIA's Nemotron-Cascade-2 models.
- Records
- 1,940,375
- Format
- PARQUET
- Downloads
- 3
WideSearch Agentic Information-Seeking Benchmark
Benchmark for evaluating LLM-driven agents on broad information-seeking tasks. Each instance includes a query, structured evaluation criteria, and language metadata for measuring agent search performance.
- Records
- 200
- Format
- PARQUET
- Downloads
- 7
English Tweets Labeled by Emotion
English Twitter messages labeled with six basic emotions: anger, fear, joy, love, sadness, and surprise. Commonly used benchmark for emotion classification and sentiment analysis research.
- Records
- 436,809
- Format
- PARQUET
- Downloads
- 3
OpenMath GSM8K Masked Grade-School Math
Grade-school math word problems with masked and full reference solutions plus expected answers, designed to support synthetic solution generation.
- Records
- 7,473
- Format
- PARQUET
- Downloads
- 2
MiniPile Deduplicated Text Corpus
Compact subset of the deduplicated Pile corpus containing approximately 1 million documents, designed for efficient language-model pretraining and experimentation.
- Records
- 1,010,500
- Format
- PARQUET
- Downloads
- 3
PII Masking Dataset
Text samples annotated for personally identifiable information detection and masking, with paired source and target text, span labels, and token-level annotations across multiple languages.
- Records
- 225,405
- Format
- PARQUET
- Downloads
- 4
Stack Overflow Posts Markdown
before June 14, 2023
Questions and answers with scores, tags, view counts, creation and edit timestamps. Includes full post metadata for each record.
- Records
- 58,329,355
- Format
- PARQUET
- Downloads
- 4
Waveform-5000 Classification Benchmark
Synthetic 3-class waveform classification dataset with 40 numeric attributes, including 21 informative features and 19 noise dimensions. Originally from Breiman et al. 1984.
- Records
- 5,000
- Format
- PARQUET
- Downloads
- 6
Nemotron Content Safety Audio Dataset
English audio files covering 23 violation categories designed for evaluating multimodal content-safety guardrails. Extends NVIDIA's Aegis 2.0 benchmark into the audio modality with adversarial and safety-critical examples.
- Records
- 1,928
- Format
- PARQUET
- Downloads
- 4
Hindawi Arabic Books Section-Level NLP Dataset
Cleaned, section-level Arabic text from Hindawi.org books spanning literature, philosophy, history, and science, prepared for Arabic NLP training and research.
- Records
- 52,830
- Format
- PARQUET
- Downloads
- 2
MInDS-14 Multilingual Spoken Intent Detection
Spoken intent detection benchmark with e-banking intents across 14 language varieties, comprising audio recordings and transcriptions for speech understanding evaluation and multilingual ASR/NLU fine-tuning.
- Records
- 16,336
- Format
- PARQUET
- Downloads
- 3
AG News Topic Classification Benchmark
News articles classified into four topics: World, Sports, Business, and Science/Technology. Standard benchmark for evaluating text classification models and NLP systems.
- Records
- 127,600
- Format
- PARQUET
- Downloads
- 3
Go-Code-Large Corpus
Curated collection of Go source code samples in JSONL format, designed for LLM pretraining, code generation fine-tuning, and static analysis research on cloud-native and backend systems.
- Records
- 316,427
- Format
- PARQUET
- Downloads
- 3
TaskTrove Agentic Tasks
Collection of agentic tasks aggregated from 100+ sources including SWE-Smith, R2EGym, and SWE-Re-Bench, designed for reinforcement learning and supervised fine-tuning of AI agents.
- Records
- 17,191
- Format
- PARQUET
- Downloads
- 3
Bhasha SFT Multilingual Instruction-Response Pairs
Instruction-response pairs in Hindi, Bengali, Gujarati, and English combining human-annotated and synthetic data for supervised fine-tuning of multilingual language models.
- Records
- 18,139,035
- Format
- PARQUET
- Downloads
- 1
SciCode Scientific Domain Code
Scientific source code from GitHub spanning biology, chemistry, materials science, physics, and 174 additional domains. Covers 155,855 code repositories with 1.1 billion+ lines of code across approximately 115 GB.
- Records
- 155,855
- Format
- PARQUET
- Downloads
- 1
HPDv3 Human Preference Dataset
Text-to-image preference dataset comprising 1.08M text-image pairs and 1.17M pairwise human annotations. Used to train HPSv3 reward models for evaluating generated images.
- Records
- 1,154,324
- Format
- PARQUET
- Downloads
- 3
LongCoT Long-Horizon Reasoning Benchmark
Benchmark dataset for evaluating sustained chain-of-thought reasoning across logic, computer science, chemistry, chess, and mathematics domains.
- Records
- 5,004
- Format
- PARQUET
- Downloads
- 7
UGMathBench Undergraduate Math Reasoning Benchmark
Undergraduate-level math problems spanning 16 subjects and 111 topics with 10 answer types. Designed for evaluating LLM mathematical reasoning capabilities.
- Records
- 5,061
- Format
- PARQUET
- Downloads
- 2
Common Crawl Creative Commons Fine Multilingual Corpus
Filtered Creative Commons web corpus from Common Crawl, intersected with FineWeb for quality. Covers English, German, French, Dutch, Spanish, Italian, Afrikaans, and Frisian.
- Records
- 75,055,472
- Format
- PARQUET
- Downloads
- 3
MASSIVE Multilingual NLU Dataset
Parallel natural language understanding benchmark with utterances annotated across 60 intents and 55 slot types, covering 51 languages. Localized from voice assistant interactions.
- Records
- 2,560,755
- Format
- PARQUET
- Downloads
- 2
Salad-Data LLM Safety Evaluation Benchmark
Aggregated safety questions for LLM red-teaming and jailbreak evaluation, sourced from HH-RLHF, AdvBench, ToxicChat, GPTFuzzer, and GPT-3.5 self-instructed prompts.
- Records
- 30,358
- Format
- PARQUET
- Downloads
- 3