Datasets:

Modalities:
Text
Formats:
parquet
Languages:
Japanese
Libraries:
Datasets
pandas
License:
teruo6939 commited on
Commit
44d9716
·
verified ·
1 Parent(s): 0de2cb3

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +27 -14
README.md CHANGED
@@ -115,22 +115,13 @@ configs:
115
  path: v1.1-extended-answers/validation-*
116
  ---
117
 
118
- # JEMHopQA
119
 
120
- > JEMHopQA (Japanese Explainable Multi-hop Question Answering) is a Japanese multi-hop QA dataset that can evaluate internal reasoning. It is a task that takes a question as input and generates an answer and derivations. Derivations are a set of derivation steps and is a semi-structured representation of relationships between entities. This dataset contains both compositional (linking information from two Wikipedia articles) and comparison (comparing information from two Wikipedia articles) questions.
121
-
122
- Source: [aiishii/JEMHopQA on GitHub](https://github.com/aiishii/JEMHopQA)
123
 
 
124
 
125
- ### Data Fields
126
-
127
- - `qid` (`str`): Unique identifier for each entry in the dataset.
128
- - `type` (`str`): The category of the question ("comparison" or "compositional").
129
- - `question` (`str`): The text of the question.
130
- - `answer` (`str`): The correct answer to the question.
131
- - `derivations` (`dict[str, list[str]]`): Knowledge triples for reasoning used to arrive at the answer.
132
- - `page_ids` (`list[str]`): Identifiers for related Wikipedia pages.
133
- - `time_dependent` (`bool`): Indicates whether the question/answer is time-sensitive.
134
 
135
  ### Licensing Information
136
 
@@ -159,4 +150,26 @@ Source: [aiishii/JEMHopQA on GitHub](https://github.com/aiishii/JEMHopQA)
159
  url = "https://aclanthology.org/2024.lrec-main.831",
160
  pages = "9515--9525",
161
  }
162
- ```
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
115
  path: v1.1-extended-answers/validation-*
116
  ---
117
 
118
+ 評価スコアの再現性確保と SB Intuitions 修正版の公開用クローン
119
 
120
+ ソース: [aiishii/JEMHopQA on GitHub](https://github.com/aiishii/JEMHopQA)
 
 
121
 
122
+ ## JEMHopQA
123
 
124
+ > JEMHopQA (Japanese Explainable Multi-hop Question Answering) is a Japanese multi-hop QA dataset that can evaluate internal reasoning. It is a task that takes a question as input and generates an answer and derivations. Derivations are a set of derivation steps and is a semi-structured representation of relationships between entities. This dataset contains both compositional (linking information from two Wikipedia articles) and comparison (comparing information from two Wikipedia articles) questions.
 
 
 
 
 
 
 
 
125
 
126
  ### Licensing Information
127
 
 
150
  url = "https://aclanthology.org/2024.lrec-main.831",
151
  pages = "9515--9525",
152
  }
153
+ ```
154
+
155
+ ## Subsets
156
+
157
+ ### v1, v1.1
158
+
159
+ - v1: [https://github.com/aiishii/JEMHopQA/tree/main/corpus](https://github.com/aiishii/JEMHopQA/tree/main/corpus)
160
+ - v1.1: [https://github.com/aiishii/JEMHopQA/tree/main/corpus_ver1.1](https://github.com/aiishii/JEMHopQA/tree/main/corpus_ver1.1)
161
+
162
+
163
+ - `qid` (`str`): Unique identifier for each entry in the dataset.
164
+ - `type` (`str`): The category of the question ("comparison" or "compositional").
165
+ - `question` (`str`): The text of the question.
166
+ - `answer` (`str`): The correct answer to the question.
167
+ - `derivations` (`dict[str, list[str]]`): Knowledge triples for reasoning used to arrive at the answer.
168
+ - `page_ids` (`list[str]`): Identifiers for related Wikipedia pages.
169
+ - `time_dependent` (`bool`): Indicates whether the question/answer is time-sensitive.
170
+
171
+ ### v1.1-extended-answers
172
+
173
+ v1.1 の answer に別解を加え、`answers` (`list[str]`) に拡張したもの
174
+
175
+ - e.g., `"カリフォルニア州クパチーノ"` -> `["カリフォルニア州クパチーノ", "アメリカ合衆国カリフォルニア州クパチーノ", "アメリカ合衆国カリフォルニア州クパティーノ", "カリフォルニア州クパティーノ"]`