llama-anon
commited on
Commit
•
7b9bc16
1
Parent(s):
2ecdd9a
Update README.md
Browse filesAdd HuggingFaceH4 benchmark metrics
README.md
CHANGED
@@ -3,6 +3,7 @@ language:
|
|
3 |
- en
|
4 |
datasets:
|
5 |
- garage-bAInd/OpenPlatypus
|
|
|
6 |
---
|
7 |
|
8 |
# Platypus2-70B-instruct
|
@@ -15,11 +16,11 @@ Platypus-70B-instruct is a merge of [`garage-bAInd/Platypus2-70B`](https://huggi
|
|
15 |
|
16 |
| Metric | Value |
|
17 |
|-----------------------|-------|
|
18 |
-
| MMLU (5-shot) |
|
19 |
-
| ARC (25-shot) |
|
20 |
-
| HellaSwag (10-shot) |
|
21 |
-
| TruthfulQA (0-shot) |
|
22 |
-
| Avg. |
|
23 |
|
24 |
We use state-of-the-art [Language Model Evaluation Harness](https://github.com/EleutherAI/lm-evaluation-harness) to run the benchmark tests above, using the same version as the HuggingFace LLM Leaderboard. Please see below for detailed instructions on reproducing benchmark results.
|
25 |
|
|
|
3 |
- en
|
4 |
datasets:
|
5 |
- garage-bAInd/OpenPlatypus
|
6 |
+
license: agpl-3.0
|
7 |
---
|
8 |
|
9 |
# Platypus2-70B-instruct
|
|
|
16 |
|
17 |
| Metric | Value |
|
18 |
|-----------------------|-------|
|
19 |
+
| MMLU (5-shot) | 70.48 |
|
20 |
+
| ARC (25-shot) | 71.84 |
|
21 |
+
| HellaSwag (10-shot) | 87.94 |
|
22 |
+
| TruthfulQA (0-shot) | 62.26 |
|
23 |
+
| Avg. | 73.13 |
|
24 |
|
25 |
We use state-of-the-art [Language Model Evaluation Harness](https://github.com/EleutherAI/lm-evaluation-harness) to run the benchmark tests above, using the same version as the HuggingFace LLM Leaderboard. Please see below for detailed instructions on reproducing benchmark results.
|
26 |
|