Dataset Card for MIRACL This is a reformatting of the MIRACL dataset used to train the BGE M3 model. See the full BGE M3 dataset in Shitao/bge m3 data. Dataset Subsets ... triplet subset Columns: "anchor", "positive", "negative" Column types: str , str , str Examples: Collection strategy: Reading the MIRACL jsonl files in Shitao/bge m3 data and taking only the first positive and first negative. Deduplified: No ... triplet 100 subset Columns: "anchor", "positive", "negative 1", ..., "negative 100" Column types: str , ..., str Examples: Collection strategy: Reading the MIRACL jsonl files in Shitao/bge m3 data, using all positives and the 100 negatives. Deduplified: No ... triplet all subset Columns: "anchor", "positive", "negative" Column types: str , str , str Examples: Collection strategy: Reading the MIRACL jsonl files in Shitao/bge m3 data and taking each negative, but making a separate sample with each of the negatives. Deduplified: No
We use cookies for essential functionality and analytics. You can accept or reject analytics cookies.Cookie policy