Commit
•
5f5a78b
1
Parent(s):
d048eee
Update README.md
Browse files
README.md
CHANGED
|
@@ -2,6 +2,18 @@
|
|
| 2 |
license: cc-by-nc-4.0
|
| 3 |
datasets:
|
| 4 |
- HuggingFaceH4/ultrafeedback_binarized
|
|
|
|
|
|
|
| 5 |
---
|
| 6 |
|
| 7 |
-
Trained for one epoch on ultrafeedback_binarized using cDPO. Evaluation pending.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2 |
license: cc-by-nc-4.0
|
| 3 |
datasets:
|
| 4 |
- HuggingFaceH4/ultrafeedback_binarized
|
| 5 |
+
language:
|
| 6 |
+
- en
|
| 7 |
---
|
| 8 |
|
| 9 |
+
Trained for one epoch on ultrafeedback_binarized using cDPO. Evaluation pending.
|
| 10 |
+
|
| 11 |
+
Some initial benchmark results:
|
| 12 |
+
| Task |Version| Metric |Value | |Stderr|
|
| 13 |
+
|---------|------:|--------|-----:|---|-----:|
|
| 14 |
+
|hellaswag| 0|acc |0.6621|± |0.0047|
|
| 15 |
+
| | |acc_norm|0.8525|± |0.0035|
|
| 16 |
+
|arc_challenge| 0|acc |0.6348|± |0.0141|
|
| 17 |
+
| | |acc_norm|0.6698|± |0.0137|
|
| 18 |
+
|winogrande| 0|acc |0.7861|± |0.0115|
|
| 19 |
+
|gsm8k| 0|acc |0.5694|± |0.0136|
|