Instructions to use knowingly-wrong/realistic-scenarios with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use knowingly-wrong/realistic-scenarios with PEFT:
Task type is invalid.
- Notebooks
- Google Colab
- Kaggle
Upload saturn_moons/meta.json with huggingface_hub
Browse files- saturn_moons/meta.json +4 -2
saturn_moons/meta.json
CHANGED
|
@@ -194,8 +194,10 @@
|
|
| 194 |
},
|
| 195 |
"assertion": {
|
| 196 |
"false_rate": 0.0,
|
| 197 |
-
"true_rate":
|
| 198 |
-
"n": 50
|
|
|
|
|
|
|
| 199 |
}
|
| 200 |
}
|
| 201 |
}
|
|
|
|
| 194 |
},
|
| 195 |
"assertion": {
|
| 196 |
"false_rate": 0.0,
|
| 197 |
+
"true_rate": 96.0,
|
| 198 |
+
"n": 50,
|
| 199 |
+
"method": "LLM judge (claude-sonnet-5): asserts true/false claim as currently correct",
|
| 200 |
+
"judge": "anthropic/claude-sonnet-5"
|
| 201 |
}
|
| 202 |
}
|
| 203 |
}
|