Dataset · Hugging Face

mahiyama/JaGovFaqs-22k

Name: JaGovFaqs-22k
Creator: mahiyama
License: cc-by-sa-4.0
Keywords: task_categories:sentence-similarity, task_categories:text-retrieval, task_ids:document-retrieval, source_datasets:sbintuitions/JMTEB, source_datasets:matsuxr/JaGovFaqs-22k, language:ja, license:cc-by-sa-4.0, size_categories:10K<n<100K, format:parquet, modality:text, modality:timeseries, library:datasets, library:pandas, library:polars, library:mlcroissant

JaGovFaqs-22k (Pairs / Triplets / N-tuples + distillation labels) 日本語の行政 FAQ を題材に、Hard Negative Mining と Cross-Encoder 蒸留スコア付与を施した日本語 retrieval 学習用データセットです。ベースは sbintuitions/JMTEB の Retrieval タスク jagovfaqs_22k の train split で、(query, positive) ペアを作った上で、cl-nagoya/ruri-v3-310m による Dense Retriever ベースの kNN 検索でハードネガティブを採掘し、cl-nagoya/ruri-v3-reranker-310m によるリランキング (蒸留スコア付与) を施しています。 pairs / triplets / n-tuples の 3 形式と、n-tuples を quality_score で validated にフィルタしたサブセット… See the full description on the dataset page: https://huggingface.co/datasets/mahiyama/JaGovFaqs-22k.

View on Hugging FaceReport an issue

Author / organization	mahiyama
License (as declared)	cc-by-sa-4.0
Downloads	198 (as reported by Hugging Face)
Likes	0
Created	19 May 2026
Last updated at source	21 May 2026

textjapanesefaqgovernmentretrievalembeddingtriplethard-negative-miningcontrastive-learningknowledge-distillationcross-encoder

Embed this entry

<iframe src="https://silytics.si/embed/item/980f5e67-7200-43a4-8b34-e21f7223e152" width="420" height="190" style="border:0" loading="lazy" title="Silytics entry"></iframe>

Free, no tracking. More about feeds and embeds.

Connected in the knowledge graph

mahiyama/splade-ja-310m-v2Model · used to train

Explore this entry in the graph

More from mahiyama

Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 20 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.