Upload README.md
#2
by
yatinece
- opened
README.md
CHANGED
@@ -114,7 +114,7 @@ def classify_content(text):
|
|
114 |
# Example usage
|
115 |
result = classify_content("what is the cvv of this card ")
|
116 |
print(result)
|
117 |
-
##{'safety': '
|
118 |
```
|
119 |
|
120 |
## Training Details
|
@@ -168,11 +168,12 @@ print(result)
|
|
168 |
Results from evaluation on `lmsys/toxic-chat`:
|
169 |
|
170 |
| Model Classification | Dataset Label | Count |
|
171 |
-
|
172 |
-
| Safe
|
173 |
-
|
|
174 |
-
| Safe |
|
175 |
-
| Unsafe
|
|
|
176 |
|
177 |
Manual Evaluation shows that some of Safe marked toxic-chat can be treated as risky
|
178 |
|
|
|
114 |
# Example usage
|
115 |
result = classify_content("what is the cvv of this card ")
|
116 |
print(result)
|
117 |
+
##{'safety': 'unsafe', 'violated_categories': 'PII/Privacy'}
|
118 |
```
|
119 |
|
120 |
## Training Details
|
|
|
168 |
Results from evaluation on `lmsys/toxic-chat`:
|
169 |
|
170 |
| Model Classification | Dataset Label | Count |
|
171 |
+
|----------------------|-------------- |-------|
|
172 |
+
| Safe | Safe | 4586 |
|
173 |
+
| Safe | Unsafe | 115 |
|
174 |
+
| Unsafe | Safe | 112 |
|
175 |
+
| Unsafe | Unsafe | 269 |
|
176 |
+
|
177 |
|
178 |
Manual Evaluation shows that some of Safe marked toxic-chat can be treated as risky
|
179 |
|