jorses
commited on
Commit
·
94afc9a
1
Parent(s):
7e65c3f
Update numbers in readme
Browse files
README.md
CHANGED
@@ -708,13 +708,12 @@ configs:
|
|
708 |
# 💾🏋️💾 DataBench 💾🏋️💾
|
709 |
|
710 |
This repository contains the original 80 datasets used for the paper [Question Answering over Tabular Data with DataBench:
|
711 |
-
A Large-Scale Empirical Evaluation of LLMs](https://huggingface.co/datasets/cardiffnlp/databench/resolve/main/Databench-LREC-Coling-2024.pdf) which appeared in LREC-COLING 2024.
|
712 |
|
713 |
Large Language Models (LLMs) are showing emerging abilities, and one of the latest recognized ones is tabular
|
714 |
reasoning in question answering on tabular data. Although there are some available datasets to assess question
|
715 |
answering systems on tabular data, they are not large and diverse enough to evaluate this new ability of LLMs.
|
716 |
-
To this end, we provide a corpus of
|
717 |
-
|
718 |
|
719 |
## Usage
|
720 |
|
|
|
708 |
# 💾🏋️💾 DataBench 💾🏋️💾
|
709 |
|
710 |
This repository contains the original 80 datasets used for the paper [Question Answering over Tabular Data with DataBench:
|
711 |
+
A Large-Scale Empirical Evaluation of LLMs](https://huggingface.co/datasets/cardiffnlp/databench/resolve/main/Databench-LREC-Coling-2024.pdf) which appeared in LREC-COLING 2024 and the associated SemEval 2025 Task 8 competition.
|
712 |
|
713 |
Large Language Models (LLMs) are showing emerging abilities, and one of the latest recognized ones is tabular
|
714 |
reasoning in question answering on tabular data. Although there are some available datasets to assess question
|
715 |
answering systems on tabular data, they are not large and diverse enough to evaluate this new ability of LLMs.
|
716 |
+
To this end, we provide a corpus of 80 real-world datasets, comprising a total of 3,708,884 rows, 2,006 columns, and 1,822 questions to evaluate models for the task of question answering (QA) over tabular data.
|
|
|
717 |
|
718 |
## Usage
|
719 |
|