Feature Extraction

Outputs raw embeddings for downstream tasks like clustering, deduplication, or feeding a classifier. Closely related to sentence similarity but exposes the vectors directly.