Update README.md
Browse files
README.md
CHANGED
|
@@ -17,7 +17,10 @@ In a pre-processing step, we split documents in into chunks of exactly 1024 toke
|
|
| 17 |
**Guidance on Responsible Use:**
|
| 18 |
|
| 19 |
In the paper, we document various types of bias that are present in the quality ratings (biases related to domains, topics, social roles, regions and languages - see Section 6 of the paper).
|
| 20 |
-
Hence, be aware that data selection with QuRating could have unintended and harmful effects on the language model that is being trained.
|
|
|
|
|
|
|
|
|
|
| 21 |
|
| 22 |
**Citation:**
|
| 23 |
```
|
|
|
|
| 17 |
**Guidance on Responsible Use:**
|
| 18 |
|
| 19 |
In the paper, we document various types of bias that are present in the quality ratings (biases related to domains, topics, social roles, regions and languages - see Section 6 of the paper).
|
| 20 |
+
Hence, be aware that data selection with QuRating could have unintended and harmful effects on the language model that is being trained.
|
| 21 |
+
We strongly recommend a comprehensive evaluation of the language model for these and other types of bias, particularly before real-world deployment.
|
| 22 |
+
We hope that releasing the data/models can facilitate future research aimed at uncovering and mitigating such biases.
|
| 23 |
+
Note that the quality ratings do not measure the social or literary value of a text and should *not* be used for textual or demographic studies.
|
| 24 |
|
| 25 |
**Citation:**
|
| 26 |
```
|