boyiwei commited on
Commit
b33c0f2
·
verified ·
1 Parent(s): 2546fe9

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +9 -9
README.md CHANGED
@@ -32,23 +32,23 @@ configs:
32
 
33
  CoTaEval Dataset is used to evaluate the feasibility and the side effects of copyright takedown methods for language models. The dataset has two domains: News and Books.
34
  For News, it has three subsets: ``news_for_unlearning`` (for unlearning use), ``news_infringement``(for infringement evaluation), and ``news_utility`` (for utility evaluation);
35
- For Books, it has two subsets: ``books_infringement`` (for infringement evaluation), ``books_utility`` (for utility evaluation). The structure of the dataset are shown below:
36
  - CoTaEval
37
  - News
38
  - news_for_unlearning
39
- - forget_set
40
- - retain_set
41
  - news_infringement
42
- - blocklisted
43
  - news_utility
44
- - blocklisted
45
- - in_domain
46
  - Books
47
  - books_infringement
48
- - blocklisted
49
  - books_utility
50
- - blocklisted
51
- - in_domain
52
 
53
  # Usage
54
 
 
32
 
33
  CoTaEval Dataset is used to evaluate the feasibility and the side effects of copyright takedown methods for language models. The dataset has two domains: News and Books.
34
  For News, it has three subsets: ``news_for_unlearning`` (for unlearning use), ``news_infringement``(for infringement evaluation), and ``news_utility`` (for utility evaluation);
35
+ For Books, it has two subsets: ``books_infringement`` (for infringement evaluation), ``books_utility`` (for utility evaluation). The structure of the dataset is shown below:
36
  - CoTaEval
37
  - News
38
  - news_for_unlearning
39
+ - forget_set (1k rows)
40
+ - retain_set (1k rows)
41
  - news_infringement
42
+ - blocklisted (1k rows)
43
  - news_utility
44
+ - blocklisted (500 rows)
45
+ - in_domain (500 rows)
46
  - Books
47
  - books_infringement
48
+ - blocklisted (500 rows)
49
  - books_utility
50
+ - blocklisted (500 rows, here we only use first 200 rows for utility evaluation)
51
+ - in_domain (200 rows)
52
 
53
  # Usage
54