MichaelYang-lyx commited on
Commit
ba767c4
·
verified ·
1 Parent(s): 6ddc833

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +17 -3
README.md CHANGED
@@ -12,10 +12,27 @@ tags:
12
  - document-understanding
13
  - benchmark
14
  pretty_name: AIDABench
 
 
 
 
 
 
 
 
 
 
 
 
 
15
  ---
16
 
17
  # Dataset Card for AIDABench
18
 
 
 
 
 
19
  ## Dataset Summary
20
 
21
  **AIDABench** is a benchmark for evaluating AI systems on **end-to-end data analytics over real-world documents**. It contains **600+** diverse analytical tasks grounded in realistic scenarios and spans heterogeneous data sources such as **spreadsheets, databases, financial reports, and operational records**. Tasks are designed to be challenging, often requiring multi-step reasoning and tool use to complete reliably.
@@ -94,9 +111,6 @@ AIDABench is intended for:
94
  - The benchmark is designed for tool-augmented settings; purely text-only inference may underperform due to the need for code execution and file manipulation.
95
  - Automated evaluation relies on LLM judges, which introduces additional compute cost and (small) scoring variance depending on settings.
96
 
97
- ## Links
98
-
99
- - **GitHub Repository**: [https://github.com/MichaelYang-lyx/AIDABench](https://github.com/MichaelYang-lyx/AIDABench)
100
 
101
  ## Citation
102
 
 
12
  - document-understanding
13
  - benchmark
14
  pretty_name: AIDABench
15
+ configs:
16
+ - config_name: qa
17
+ data_files:
18
+ - split: test
19
+ path: QA/QA.jsonl
20
+ - config_name: data_visualization
21
+ data_files:
22
+ - split: test
23
+ path: data_visualization/data_visualization.jsonl
24
+ - config_name: file_generation
25
+ data_files:
26
+ - split: test
27
+ path: file_generation/file_generation.jsonl
28
  ---
29
 
30
  # Dataset Card for AIDABench
31
 
32
+ ## Links
33
+ - [Paper (arXiv)](https://arxiv.org/abs/2603.15636)
34
+ - [GitHub Repository](https://github.com/MichaelYang-lyx/AIDABench)
35
+
36
  ## Dataset Summary
37
 
38
  **AIDABench** is a benchmark for evaluating AI systems on **end-to-end data analytics over real-world documents**. It contains **600+** diverse analytical tasks grounded in realistic scenarios and spans heterogeneous data sources such as **spreadsheets, databases, financial reports, and operational records**. Tasks are designed to be challenging, often requiring multi-step reasoning and tool use to complete reliably.
 
111
  - The benchmark is designed for tool-augmented settings; purely text-only inference may underperform due to the need for code execution and file manipulation.
112
  - Automated evaluation relies on LLM judges, which introduces additional compute cost and (small) scoring variance depending on settings.
113
 
 
 
 
114
 
115
  ## Citation
116