Datasets:
Premise stringlengths 18 89 | Hypothesis stringlengths 21 77 | Label stringclasses 3
values |
|---|---|---|
They got married on Saturday. | They got married before Friday. | contradiction |
They got married on Sunday. | They got married after Monday. | contradiction |
They got married on Thursday. | They got married after Saturday. | contradiction |
They got married on Sunday. | They got married after Monday. | contradiction |
They got married on Friday. | They got married before Friday. | contradiction |
On Wednesday, they got married. | Before Friday, they got married. | entailment |
On Saturday, they got married. | Before Friday, they got married. | contradiction |
On Thursday, they got married. | After Monday, they got married. | entailment |
On Monday, they got married. | After Thursday, they got married. | contradiction |
On Sunday, they got married. | After Sunday, they got married. | contradiction |
They got married after Saturday. | They got married before Saturday. | contradiction |
They got married before Friday. | They got married before Wednesday. | neutral |
They got married before Wednesday. | They got married after Tuesday. | neutral |
They got married after Sunday. | They got married after Tuesday. | neutral |
They got married after Tuesday. | They got married before Tuesday. | contradiction |
Before Monday, they got married. | Before Wednesday, they got married. | entailment |
After Wednesday, they got married. | Before Thursday, they got married. | neutral |
After Wednesday, they got married. | After Thursday, they got married. | neutral |
Before Saturday, they got married. | After Wednesday, they got married. | neutral |
After Tuesday, they got married. | After Wednesday, they got married. | neutral |
They got married in Jul. | They got married before December. | entailment |
They got married in August. | They got married after Oct. | contradiction |
They got married in Aug. | They got married after July. | entailment |
They got married in Nov. | They got married after October. | entailment |
They got married in Apr. | They got married after March. | entailment |
In Jan, they got married. | Before May, they got married. | entailment |
In September, they got married. | After Mar, they got married. | entailment |
In Jul, they got married. | Before August, they got married. | entailment |
In December, they got married. | After Sep, they got married. | entailment |
In August, they got married. | Before Dec, they got married. | entailment |
They got married after April. | They got married before Aug. | neutral |
They got married before December. | They got married after Nov. | neutral |
They got married before May. | They got married after April. | neutral |
They got married after May. | They got married before March. | contradiction |
They got married before November. | They got married before Jun. | neutral |
Before Nov, they got married. | Before November, they got married. | neutral |
Before April, they got married. | Before May, they got married. | entailment |
After Aug, they got married. | After May, they got married. | entailment |
After April, they got married. | Before Mar, they got married. | contradiction |
After Dec, they got married. | After November, they got married. | entailment |
They got married in 1981. | They got married after 1924. | entailment |
They got married in 1950. | They got married before 1913. | contradiction |
They got married in 1968. | They got married after 1922. | entailment |
They got married in 1957. | They got married before 1951. | contradiction |
They got married in 1911. | They got married after 1949. | contradiction |
In 1937, they got married. | Before 1969, they got married. | entailment |
In 1942, they got married. | After 1921, they got married. | entailment |
In 1973, they got married. | Before 1917, they got married. | contradiction |
In 1966, they got married. | After 1979, they got married. | contradiction |
In 1998, they got married. | After 1942, they got married. | entailment |
They got married after 1973. | They got married before 1917. | contradiction |
They got married after 1954. | They got married before 1950. | contradiction |
They got married after 1985. | They got married before 1915. | contradiction |
They got married before 1977. | They got married before 1957. | neutral |
They got married after 1992. | They got married before 1925. | contradiction |
After 1927, they got married. | After 1914, they got married. | entailment |
Before 1948, they got married. | After 1959, they got married. | contradiction |
Before 1941, they got married. | Before 1990, they got married. | entailment |
After 1990, they got married. | Before 1948, they got married. | contradiction |
Before 1930, they got married. | Before 1911, they got married. | neutral |
They got married in September. | They got married before September. | contradiction |
They got married in August. | They got married before June. | contradiction |
They got married in August. | They got married before October. | entailment |
They got married in February. | They got married after January. | entailment |
They got married in November. | They got married before October. | contradiction |
In September, they got married. | After May, they got married. | entailment |
In April, they got married. | After May, they got married. | contradiction |
In September, they got married. | Before September, they got married. | contradiction |
In September, they got married. | Before November, they got married. | entailment |
In February, they got married. | Before April, they got married. | entailment |
They got married after February. | They got married after July. | neutral |
They got married before August. | They got married before October. | entailment |
They got married after October. | They got married before November. | neutral |
They got married before November. | They got married before August. | neutral |
They got married before February. | They got married after August. | contradiction |
Before November, they got married. | After November, they got married. | contradiction |
Before November, they got married. | Before December, they got married. | entailment |
After July, they got married. | Before September, they got married. | neutral |
Before October, they got married. | Before September, they got married. | neutral |
After August, they got married. | Before August, they got married. | contradiction |
They got married in Oct. | They got married after Jun. | entailment |
They got married in Dec. | They got married after Jul. | entailment |
They got married in Nov. | They got married after Dec. | contradiction |
They got married in Nov. | They got married before Jul. | contradiction |
They got married in Feb. | They got married after Mar. | contradiction |
In Apr, they got married. | Before May, they got married. | entailment |
In Apr, they got married. | After May, they got married. | contradiction |
In Jul, they got married. | Before Jan, they got married. | contradiction |
In Sep, they got married. | After Aug, they got married. | entailment |
In Aug, they got married. | Before Aug, they got married. | contradiction |
They got married after Feb. | They got married after May. | neutral |
They got married after Feb. | They got married after Aug. | neutral |
They got married after Dec. | They got married before Nov. | contradiction |
They got married before Mar. | They got married before Jun. | entailment |
They got married before Sep. | They got married after Oct. | contradiction |
Before Jun, they got married. | After Mar, they got married. | neutral |
Before Aug, they got married. | Before Apr, they got married. | neutral |
Before Sep, they got married. | Before May, they got married. | neutral |
After Sep, they got married. | After Nov, they got married. | neutral |
Before Oct, they got married. | Before Jul, they got married. | neutral |
End of preview. Expand in Data Studio
tasksource/temporal-nli(複製)
データの内容
時間の前後関係、期間、異なる単位の時間の大きさの関係を調べる、3 種類の NLI の challenge set のデータ(Thukral ら 2021)。1 行は前提(Premise)と仮説(Hypothesis)の組で、正解が Label に入る。論文の著者らが作った challenge set である。
データの概要
| 項目 | 値 |
|---|---|
| 行数 | 98,240(1 構成・2 分割の合計) |
| 列数 | 3 |
| 容量 | 10.77 MB(ファイル 4 本) |
| ファイルの形式 | csv 2 |
| 公開 | public |
構成と分割
| 構成(config) | 分割(split)ごとの行数 | 列数 |
|---|---|---|
default |
train 72,720、test 25,520 |
3 |
列(構成 default・分割 train)
| 列 | 型 | 先頭の行の値 |
|---|---|---|
Premise |
string |
They got married on Saturday. |
Hypothesis |
string |
They got married before Friday. |
Label |
string |
contradiction |
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 にある。
複製について
tasksource/temporal-nli をリビジョン d5cedbbdb9f1e7591569ebaf7cf1dd238f0b624b 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/temporal-nli
- 複製したリビジョン:
d5cedbbdb9f1e7591569ebaf7cf1dd238f0b624b(2023-03-24 10:30:25+00:00) - 元のライセンス表記:
apache-2.0
論文
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: A(許諾型のライセンスだけで、疑義の注記が無い)。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) | apache-2.0 / license_use=commercial | 確認済 | license_use=commercial / [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) | Apache-2.0 | 確認済 | [Jev 元データ source] source_annotation_converted(Choice/Noul/Score recast; soft t |
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。
@inproceedings{thukral-etal-2021-probing,
title = "Probing Language Models for Understanding of Temporal Expressions",
author = "Thukral, Shivin and
Kukreja, Kunal and
Kavouras, Christian",
booktitle = "Proceedings of the Fourth BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP",
month = nov,
year = "2021",
address = "Punta Cana, Dominican Republic",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2021.blackboxnlp-1.31",
doi = "10.18653/v1/2021.blackboxnlp-1.31",
pages = "396--406",
abstract = "We present three Natural Language Inference (NLI) challenge sets that can evaluate NLI models on their understanding of temporal expressions. More specifically, we probe these models for three temporal properties: (a) the order between points in time, (b) the duration between two points in time, (c) the relation between the magnitude of times specified in different units. We find that although large language models fine-tuned on MNLI have some basic perception of the order between points in time, at large, these models do not have a thorough understanding of the relation between temporal expressions.",
}
原文(英語)
@inproceedings{thukral-etal-2021-probing,
title = "Probing Language Models for Understanding of Temporal Expressions",
author = "Thukral, Shivin and
Kukreja, Kunal and
Kavouras, Christian",
booktitle = "Proceedings of the Fourth BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP",
month = nov,
year = "2021",
address = "Punta Cana, Dominican Republic",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2021.blackboxnlp-1.31",
doi = "10.18653/v1/2021.blackboxnlp-1.31",
pages = "396--406",
abstract = "We present three Natural Language Inference (NLI) challenge sets that can evaluate NLI models on their understanding of temporal expressions. More specifically, we probe these models for three temporal properties: (a) the order between points in time, (b) the duration between two points in time, (c) the relation between the magnitude of times specified in different units. We find that although large language models fine-tuned on MNLI have some basic perception of the order between points in time, at large, these models do not have a thorough understanding of the relation between temporal expressions.",
}
- Downloads last month
- 46