Hugging Face's logo

Datasets:
BEE-spoke-data
/
SaunaWeb-50k

Tasks:
Text Generation
Fill-Mask
Modalities:
Text
Formats:
parquet
Size:
10K<n<100K
Tags:
Libraries:
Datasets
pandas
Croissant
License:
Dataset card Viewer Files
pszemraj commited on
Commit
97676f9
•
1 Parent(s): f891fe5

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +21 -1
README.md CHANGED
@@ -31,4 +31,24 @@ task_categories:
31
 
32
 
33
 
34
- Samples from `HuggingFaceFW/fineweb` that contain the word 'sauna'
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
31
 
32
 
33
 
34
+ Samples from `HuggingFaceFW/fineweb` that contain the word 'sauna'
35
+
36
+
37
+ ## counts
38
+
39
+ GPT-4 tiktoken token count:
40
+
41
+ ```
42
+ token_count
43
+ count 50000.00000
44
+ mean 1428.86602
45
+ std 4498.05023
46
+ min 49.00000
47
+ 25% 325.00000
48
+ 50% 628.50000
49
+ 75% 1238.25000
50
+ max 147973.00000
51
+ ```
52
+
53
+
54
+ - Total count: 71.44 M tokens