Datasets:
Modalities:
Text
Formats:
json
Sub-tasks:
document-retrieval
Languages:
English
Size:
100K - 1M
Tags:
conversational-search
conversational-retrieval
information-retrieval
multi-turn
query-rewriting
benchmark
License:
Commit History
3 days ago
-
Card: point to the training negatives
Verified signature6ab17a1 -
Code link: github.com/DataScienceUIBK/MemRevise
Verified signature8cf7f37 -
Card in the paper's style: overview figure, highlights, tiers, worked example
Verified signatureedb64dd -
Each record: history + question (tier and passages belong to the question); generated subset removed; card with a worked example
Verified signaturecf29032
4 days ago
-
Shorter card; example chosen and checked by hand
Verified signatureeee9165 -
One benchmark config with train / dev / test splits (eval_set names each set); data samples in the card
Verified signature0439dc1 -
ConvRevise: benchmark, development and training questions, the three collections as JSONL, dataset card
Verified signature4723cba
8 days ago
-
remove index_orquac_titled/ (now corpora/)
Verified signaturedfea28f -
remove index_coral/ (now corpora/)
Verified signature91185bc -
remove index/ (now corpora/)
Verified signatured876b25 -
corpora/orquac
Verified signatured3ea305 -
corpora/coral
Verified signature4c7899c -
corpora/topiocqa
Verified signature675f012 -
index_orquac_titled: passage collection
Verified signature7d18ec1 -
index_coral: passage collection
Verified signature39f5bb5 -
index: passage collection
Verified signaturecc49811 -
initial commit
Verified signature358ed37