Dataset · agentlans
Cleaned English sentences from C4 and FineWeb
agentlans/high-quality-english-sentences
A spam-filtered corpus of millions of English sentences for text tasks.
What's inside
The dataset contains between 1 and 10 million individual English sentences. They are gathered from the C4 and FineWeb corpora (1 million each before filtering) and cleaned of spam and unreadable text, although spelling and grammar errors may still occur.
What you can do with it
- Fine-tuning language models and encoders for feature extraction
- Training models for semantic sentence similarity estimation and text classification
- Tasks
- Text classification, Text generation, Embeddings, Sentence similarity
- Languages
- English
- Size
- 1M–10M rows
- Downloads this month
- 1.2K
- Likes
- 41
Models trained on it
- techdotus/Ethoric-10MText generation 1.3K
- techdotus/Tokle-3MText generation 605
- mradermacher/Substitution-Cipher-Text-Eng-GGUF 306
- techdotus/Tokle-SPAB-3MText generation 155
- mradermacher/Substitution-Cipher-Alphabet-Eng-GGUF 152
- mradermacher/AutoCorrect-EN-GGUF 134
- Cipher-AI/AutoCorrect-EN 126
- mradermacher/AutoCorrect-EN-v2-GGUF 112
How to load it
from datasets import load_dataset
ds = load_dataset("agentlans/high-quality-english-sentences")