--- license: other pretty_name: tasksource/robustLR (mirror) source_datasets: - tasksource/robustLR tags: - laya - laya-finetune - mirror - domain-math - use-unclear - genre-other - label-binary - annot-rule dataset_info: features: - name: context dtype: string - name: statement dtype: string - name: label dtype: string splits: - name: train num_bytes: 2436331 num_examples: 2317 - name: test num_bytes: 3066737 num_examples: 3000 - name: dev num_bytes: 2512690 num_examples: 2317 download_size: 1330059 dataset_size: 8015758 --- # tasksource/robustLR(複製) ## データの内容 演繹推論で論理的な摂動への頑健性を評価する、Sanyal ら(2022)の RobustLR のデータ。1 行は文脈(`context`)と主張(`statement`)の組で、正誤が `label` に入る。元データの注釈を決定的な変換で作り直したものである。 ## データの概要 | 項目 | 値 | |---|---| | 行数 | 7,634(1 構成・3 分割の合計) | | 列数 | 3 | | 容量 | 1.33 MB(ファイル 5 本) | | ファイルの形式 | parquet 3 | | 公開 | public | ### 構成と分割 | 構成(config) | 分割(split)ごとの行数 | 列数 | |---|---|---:| | `default` | `train` 2,317、`test` 3,000、`dev` 2,317 | 3 | ### 列(構成 `default`・分割 `train`) | 列 | 型 | 先頭の行の値 | |---|---|---| | `context` | `string` | `Fiona is not Harry's uncle. Anne is the wife of …` | | `statement` | `string` | `Fiona is not the uncle of Gary.` | | `label` | `string` | `entailment` | 行数と列は HF のデータセットビューアの集計(元のデータセットの現在の版で数えたもの)。全データセットの一覧は [HYUKI5/dataset-catalog](https://huggingface.co/datasets/HYUKI5/dataset-catalog) にある。 ## 複製について [`tasksource/robustLR`](https://huggingface.co/datasets/tasksource/robustLR) をリビジョン `369ce16fbe4467f46916efe3730d04fb607c0391` の時点でそのまま複製したもの。中身は変えていない。判断モデル Laya(`convaiinnovations/laya`)の微調整に使う候補として、調査メモ(`Laya 学習用データセット調査`)に載せたデータを集めている。 ## 出典 - 元のデータセット: https://huggingface.co/datasets/tasksource/robustLR - 複製したリビジョン: `369ce16fbe4467f46916efe3730d04fb607c0391`(2023-06-22 14:12:37+00:00) - 元のライセンス表記: `other` ## 論文 - https://arxiv.org/abs/2205.12598 ## アノテーションの方法 元のカードにアノテーションの節が無かった。下の「調査メモの記述」と「元のカード」を参照。 ## 微調整に使うときの注意 - 調査メモの台帳での区分: D(記載なし・other・unknown・資料間の食い違い)。区分は記載の文言で機械的に付けたもので、法的な判定ではない。学習に入れる前に元のカードとライセンスの原文で確かめる。 - Laya の入力は 2,048 token まで(ModernBERT-large / mmBERT)。長い本文は先頭から黙って切られるので、切れる行は事前に除くか、要約・切り出しをしてから入れる。 - 分割は記事・文書単位で行い、同じ本文から作った別の問いを学習側と較正側に分けない。温度と閾値は学習に使っていない分割で別々に当てる。 ## 調査メモの記述 | メモ | 節 | ライセンス(記載) | 確認 | 注記 | |---|---|---|---|---| | `jev_sources_tasksource_bekko.csv` | CSV(tasksource-jev-typed-decisions) | unspecified / license_use=unspecified | 確認済 | license_use=unspecified / [Jev 元データ source] source_annotation_converted(deterministic recast; no model r | | `jev_sources_tasksource_bekko.csv` | CSV(bekko-system-one-dataset-v0) | MIT | 確認済 | [Jev 元データ source] source_annotation_or_rule_labels(eval only) | ## 元のカード(日本語訳) 以下は元のデータセットのカードの本文を日本語に訳したもの(機械翻訳)。訳に疑問があるときは、末尾の原文を確かめる。 --- # データセットカード: "robustLR" https://github.com/INK-USC/RobustLR ``` @article{sanyal2022robustlr, title={Robustlr: Evaluating robustness to logical perturbation in deductive reasoning}, author={Sanyal, Soumya and Liao, Zeyi and Ren, Xiang}, journal={arXiv preprint arXiv:2205.12598}, year={2022} } ```
原文(英語) # Dataset Card for "robustLR" https://github.com/INK-USC/RobustLR ``` @article{sanyal2022robustlr, title={Robustlr: Evaluating robustness to logical perturbation in deductive reasoning}, author={Sanyal, Soumya and Liao, Zeyi and Ren, Xiang}, journal={arXiv preprint arXiv:2205.12598}, year={2022} } ```