dardem commited on
Commit
a31a742
1 Parent(s): 7cdecd6

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +0 -5
README.md CHANGED
@@ -13,16 +13,11 @@ The ParaDetox Dataset collection was done via [Yandex.Toloka](https://toloka.yan
13
  * *Task 2:* **Content Preservation Check**: We show users the generated paraphrases along with their original variants and ask them to indicate if they have close meanings.
14
  * *Task 3:* **Toxicity Check**: Finally, we check if the workers succeeded in removing toxicity.
15
 
16
- The whole pipeline is illustrated on this schema:
17
- ![](https://github.com/skoltech-nlp/paradetox/blob/main/img/generation_pipeline_blue.jpg)
18
-
19
  All these steps were done to ensure high quality of the data and make the process of collection automated. For more details please refer to the original paper.
20
 
21
  ## ParaDetox Dataset
22
  As a result, we get paraphrases for 11,939 toxic sentences (on average 1.66 paraphrases per sentence), 19,766 paraphrases total. The whole dataset can be found [here](https://github.com/skoltech-nlp/paradetox/blob/main/paradetox/paradetox.tsv). The examples of samples from ParaDetox Dataset:
23
 
24
- ![](https://github.com/skoltech-nlp/paradetox/blob/main/img/paraphrase_example.png)
25
-
26
  In addition to all ParaDetox dataset, we also make public [samples](https://github.com/skoltech-nlp/paradetox/blob/main/paradetox/paradetox_cannot_rewrite.tsv) that were marked by annotators as "cannot rewrite" in *Task 1* of crowdsource pipeline.
27
 
28
  # Detoxification evaluation
 
13
  * *Task 2:* **Content Preservation Check**: We show users the generated paraphrases along with their original variants and ask them to indicate if they have close meanings.
14
  * *Task 3:* **Toxicity Check**: Finally, we check if the workers succeeded in removing toxicity.
15
 
 
 
 
16
  All these steps were done to ensure high quality of the data and make the process of collection automated. For more details please refer to the original paper.
17
 
18
  ## ParaDetox Dataset
19
  As a result, we get paraphrases for 11,939 toxic sentences (on average 1.66 paraphrases per sentence), 19,766 paraphrases total. The whole dataset can be found [here](https://github.com/skoltech-nlp/paradetox/blob/main/paradetox/paradetox.tsv). The examples of samples from ParaDetox Dataset:
20
 
 
 
21
  In addition to all ParaDetox dataset, we also make public [samples](https://github.com/skoltech-nlp/paradetox/blob/main/paradetox/paradetox_cannot_rewrite.tsv) that were marked by annotators as "cannot rewrite" in *Task 1* of crowdsource pipeline.
22
 
23
  # Detoxification evaluation