File size: 756 Bytes
d0d12d4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
25b0cde
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
---
pretty_name: Anonymous Data
configs:
- config_name: small
  data_files:
  - split: test
    path: "graphs/small/*.csv"
- config_name: large
  data_files:
  - split: test
    path: "graphs/large/*.csv"
- config_name: synthetic
  data_files:
  - split: test
    path: "graphs/synthetic/*.csv"
---

# Anonymous Data

This dataset contains benchmark files for evaluating language models on causal-discovery tasks.

The `small` and `synthetic` configs are available through the Hugging Face dataset viewer. The `large` config is provided as direct CSV files under `graphs/large/`; these files are large enough that automatic viewer/Croissant conversion may omit them, so the NeurIPS Croissant metadata is completed manually from the hosted CSV file layout.