Dataset · mteb
SICK-R semantic textual similarity for MTEB
mteb/sickr-sts
An English test set (1K-10K) from the MTEB benchmark for evaluating text embedding quality.
What's inside
Contains pairs of English sentences from web sources and written texts, annotated for semantic textual similarity (Semantic Textual Similarity). The dataset size ranges from 1 to 10 thousand records. The set is adapted for standard execution via the MTEB embedding evaluation library.
What you can do with it
- Test and compare text embedding models in MTEB
- Check duplicate search quality and semantic ranking
- Tasks
- Sentence similarity
- Languages
- English
- Size
- 1K–10K rows
- Downloads this month
- 22.8K
- Likes
- 6
Models trained on it
How to load it
from datasets import load_dataset
ds = load_dataset("mteb/sickr-sts")