Datasets:

MichiganNLP
/

TID-8

Tasks:

Text Classification

Modalities:

Formats:

Sub-tasks:

natural-language-inference

sentiment-analysis

hate-speech-detection

Languages:

Size:

ArXiv:

Tags:

annotation disagreement

Libraries:

License:

Dataset card Data Studio Files Files and versions

dnaihao commited on Oct 30, 2023

Commit

e12b715

·

1 Parent(s): c3c67fb

Updated readme file

Files changed (2) hide show

.gitignore +2 -1
README.md +3 -1

.gitignore CHANGED Viewed

@@ -1,3 +1,4 @@
 README-working.md
 extract_info.py
-output

 README-working.md
 extract_info.py
+output
+load_and_push.py

README.md CHANGED Viewed

@@ -894,8 +894,10 @@ configs:
 ### Dataset Summary
-TID-8 is a new benchmark focused on the task of letting models learn from data that has inherent disagreement.
 *Annotation Split*

 ### Dataset Summary
+TID-8 is a new aggregated benchmark focused on the task of letting models learn from data that has inherent disagreement proposed in [link](https://arxiv.org/pdf/2305.14663.pdf) at Findings of EMNLP 2023.
+In the paper, we focus on the inherent disagreement and let the model directly learn from data that has such disagreement.
+We provide two split for TID-8.
 *Annotation Split*