This dataset is the official benchmark for the paper "The Multilingual Curse at the Retrieval Layer:
This dataset can be used directly with Sentence Transformers to train Amharic Embedding and Rerankin