Add task category and links to paper and code
#2
by
nielsr
HF Staff
- opened
README.md
CHANGED
|
@@ -20,16 +20,20 @@ configs:
|
|
| 20 |
data_files:
|
| 21 |
- split: train
|
| 22 |
path: data/train-*
|
|
|
|
|
|
|
| 23 |
---
|
| 24 |
|
| 25 |
# Dataset Card for MMTU
|
| 26 |
|
| 27 |
## Dataset Summary
|
| 28 |
|
| 29 |
-
<!-- add link -->
|
| 30 |
MMTU: A Massive Multi-Task Table Understanding and Reasoning Benchmark
|
| 31 |
by Junjie Xing, [Yeye He](https://www.microsoft.com/en-us/research/people/yeyehe/), Mengyu Zhou, Haoyu Dong, Shi Han, Lingjiao Chen, Dongmei Zhang, [Surajit Chaudhuri](https://www.microsoft.com/en-us/research/people/surajitc/), and [H. V. Jagadish](https://web.eecs.umich.edu/~jag/).
|
| 32 |
|
|
|
|
|
|
|
|
|
|
| 33 |
This is a large-scale benchmark designed to evaluate the table reasoning capabilities of large language models (LLMs). It consists of over 30,000 questions across 25 real-world table tasks, focusing on deep understanding, reasoning, and manipulation of tabular data.
|
| 34 |
|
| 35 |
These tasks are curated from decades of computer science research and represent challenges encountered by expert users in real applications, making MMTU a rigorous test for LLMs aspiring to professional-level table understanding.
|
|
@@ -63,9 +67,6 @@ English
|
|
| 63 |
- task: The specific subtask category within the MMTU framework to which the instance belongs.
|
| 64 |
- dataset: The original source dataset from which the MMTU instance is derived.
|
| 65 |
|
| 66 |
-
|
| 67 |
## Dataset Creation
|
| 68 |
|
| 69 |
-
Please refer to Section 3.2 in the paper.
|
| 70 |
-
|
| 71 |
-
|
|
|
|
| 20 |
data_files:
|
| 21 |
- split: train
|
| 22 |
path: data/train-*
|
| 23 |
+
task_categories:
|
| 24 |
+
- table-question-answering
|
| 25 |
---
|
| 26 |
|
| 27 |
# Dataset Card for MMTU
|
| 28 |
|
| 29 |
## Dataset Summary
|
| 30 |
|
|
|
|
| 31 |
MMTU: A Massive Multi-Task Table Understanding and Reasoning Benchmark
|
| 32 |
by Junjie Xing, [Yeye He](https://www.microsoft.com/en-us/research/people/yeyehe/), Mengyu Zhou, Haoyu Dong, Shi Han, Lingjiao Chen, Dongmei Zhang, [Surajit Chaudhuri](https://www.microsoft.com/en-us/research/people/surajitc/), and [H. V. Jagadish](https://web.eecs.umich.edu/~jag/).
|
| 33 |
|
| 34 |
+
[Paper](https://huggingface.co/papers/2506.05587)
|
| 35 |
+
[Code](https://github.com/MMTU-Benchmark/MMTU)
|
| 36 |
+
|
| 37 |
This is a large-scale benchmark designed to evaluate the table reasoning capabilities of large language models (LLMs). It consists of over 30,000 questions across 25 real-world table tasks, focusing on deep understanding, reasoning, and manipulation of tabular data.
|
| 38 |
|
| 39 |
These tasks are curated from decades of computer science research and represent challenges encountered by expert users in real applications, making MMTU a rigorous test for LLMs aspiring to professional-level table understanding.
|
|
|
|
| 67 |
- task: The specific subtask category within the MMTU framework to which the instance belongs.
|
| 68 |
- dataset: The original source dataset from which the MMTU instance is derived.
|
| 69 |
|
|
|
|
| 70 |
## Dataset Creation
|
| 71 |
|
| 72 |
+
Please refer to Section 3.2 in the paper.
|
|
|
|
|
|