Update README.md
Browse files
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
|
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 |
|