Commit
·
36a0ec2
1
Parent(s):
a6e7a80
Update README.md
Browse files
README.md
CHANGED
|
@@ -61,7 +61,7 @@ other ACC: 70.04
|
|
| 61 |
|
| 62 |
social ACC: 72.41
|
| 63 |
|
| 64 |
-
**AVERAGE ACC:63.82**
|
| 65 |
|
| 66 |
## CEval (Val):
|
| 67 |
STEM acc: 61.67
|
|
@@ -74,8 +74,8 @@ Other acc: 68.35
|
|
| 74 |
|
| 75 |
Hard acc:48.03
|
| 76 |
|
| 77 |
-
**AVERAGE acc:70.27**
|
| 78 |
|
| 79 |
## GSM8K
|
| 80 |
|
| 81 |
-
**Zero-shot ACC 0.5921152388172858**
|
|
|
|
| 61 |
|
| 62 |
social ACC: 72.41
|
| 63 |
|
| 64 |
+
**AVERAGE ACC:63.82** (Outperforms / Equal to the best Mistral-7B Chat-style fine-tunes.)
|
| 65 |
|
| 66 |
## CEval (Val):
|
| 67 |
STEM acc: 61.67
|
|
|
|
| 74 |
|
| 75 |
Hard acc:48.03
|
| 76 |
|
| 77 |
+
**AVERAGE acc:70.27** (Outperforms ALL 7B models currently.)
|
| 78 |
|
| 79 |
## GSM8K
|
| 80 |
|
| 81 |
+
**Zero-shot ACC 0.5921152388172858** (Outperforms WizardMath-7B and Qwen-7B)
|