Datasets:
Modalities:
Text
Formats:
parquet
Languages:
English
Size:
100K - 1M
ArXiv:
Tags:
multi-hop-question-answering
hotpotqa
evidence-selection
question-decomposition
chain-of-thought
supervised-fine-tuning
License:
File size: 814 Bytes
7f3a1d4 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 | {
"schema_version": 1,
"upstream": {
"mode": "huggingface-datasets",
"repository": "hotpotqa/hotpot_qa",
"revision": "1908d6afbbead072334abe2965f91bd2709910ab",
"config": "distractor",
"split": "train",
"row_count": 90447
},
"annotation_patches": {
"manifest": "SOURCE_PATCHES.json",
"invalid_fact_count": 22,
"sha256": "e07d9fe7c2e8622a487753867082a21331db005408bf4f810ff710703fde019a"
},
"release": {
"row_count_per_config": 90447,
"identity_key": "source_id",
"configs": [
"structured",
"reader-sft",
"cot-reader-sft",
"paragraph-selector-sft",
"question-decomposer-sft",
"sentence-selector-sft",
"decomposed-sentence-selector-sft",
"joint-selector-reader-sft"
]
}
}
|