Adding Evaluation Results
#4
by
leaderboard-pr-bot
- opened
README.md
CHANGED
@@ -39,4 +39,17 @@ with torch.cuda.amp.autocast():
|
|
39 |
)
|
40 |
|
41 |
print(tokenizer.decode(out[0, :-1]).replace('<question>', "User:\n").replace('<answer>', '\nAssistant:\n'))
|
42 |
-
```
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
39 |
)
|
40 |
|
41 |
print(tokenizer.decode(out[0, :-1]).replace('<question>', "User:\n").replace('<answer>', '\nAssistant:\n'))
|
42 |
+
```
|
43 |
+
# [Open LLM Leaderboard Evaluation Results](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)
|
44 |
+
Detailed results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/details_OpenAssistant__galactica-6.7b-finetuned)
|
45 |
+
|
46 |
+
| Metric | Value |
|
47 |
+
|-----------------------|---------------------------|
|
48 |
+
| Avg. | 34.34 |
|
49 |
+
| ARC (25-shot) | 41.55 |
|
50 |
+
| HellaSwag (10-shot) | 51.01 |
|
51 |
+
| MMLU (5-shot) | 38.03 |
|
52 |
+
| TruthfulQA (0-shot) | 41.65 |
|
53 |
+
| Winogrande (5-shot) | 57.7 |
|
54 |
+
| GSM8K (5-shot) | 3.11 |
|
55 |
+
| DROP (3-shot) | 7.3 |
|