diffnamehard
commited on
Commit
•
fea5c1c
1
Parent(s):
18a3b2e
Update README.md
Browse files
README.md
CHANGED
@@ -5,4 +5,14 @@ tags:
|
|
5 |
---
|
6 |
This is an experimental model.
|
7 |
|
8 |
-
Finetuned on dataset [toxic-dpo-v0.1-NoWarning-alpaca](https://huggingface.co/datasets/diffnamehard/toxic-dpo-v0.1-NoWarning-alpaca) using model [Mistral-CatMacaroni-slerp-gradient](https://huggingface.co/diffnamehard/Mistral-CatMacaroni-slerp-gradient)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
5 |
---
|
6 |
This is an experimental model.
|
7 |
|
8 |
+
Finetuned on dataset [toxic-dpo-v0.1-NoWarning-alpaca](https://huggingface.co/datasets/diffnamehard/toxic-dpo-v0.1-NoWarning-alpaca) using model [Mistral-CatMacaroni-slerp-gradient](https://huggingface.co/diffnamehard/Mistral-CatMacaroni-slerp-gradient)
|
9 |
+
|
10 |
+
| Metric | Value |
|
11 |
+
| --- | --- |
|
12 |
+
| Avg. | 67.28 |
|
13 |
+
| ARC (25-shot) | 64.25 |
|
14 |
+
| HellaSwag (10-shot) | 84.09 |
|
15 |
+
| MMLU (5-shot) | 62.66 |
|
16 |
+
| TruthfulQA (0-shot) | 56.87 |
|
17 |
+
| Winogrande (5-shot) | 79.72 |
|
18 |
+
| GSM8K (5-shot) | 56.1 |
|