Update README.md
Browse files
README.md
CHANGED
|
@@ -30,7 +30,7 @@ https://AlignmentLab.ai
|
|
| 30 |
|
| 31 |
# Evaluation
|
| 32 |
|
| 33 |
-
We have evaluated
|
| 34 |
|
| 35 |
Our average performance for BigBench-Hard: 0.3753
|
| 36 |
|
|
@@ -42,10 +42,10 @@ We've done the same and have found our score averages to ~60% of the total impro
|
|
| 42 |
So we got 60% of the improvement with 6% of the data!
|
| 43 |
|
| 44 |
## BigBench-Hard Performance
|
| 45 |
-
 Evals once we receive them.
|
| 51 |
|
|
@@ -77,7 +77,7 @@ Please await our full releases for further training details.
|
|
| 77 |
year = {2023},
|
| 78 |
publisher = {HuggingFace},
|
| 79 |
journal = {HuggingFace repository},
|
| 80 |
-
howpublished = {\url{https://https://huggingface.co/Open-Orca/
|
| 81 |
}
|
| 82 |
```
|
| 83 |
```bibtex
|
|
|
|
| 30 |
|
| 31 |
# Evaluation
|
| 32 |
|
| 33 |
+
We have evaluated OpenOrca-Preview1-13B on hard reasoning tasks from BigBench-Hard and AGIEval as outlined in the Orca paper.
|
| 34 |
|
| 35 |
Our average performance for BigBench-Hard: 0.3753
|
| 36 |
|
|
|
|
| 42 |
So we got 60% of the improvement with 6% of the data!
|
| 43 |
|
| 44 |
## BigBench-Hard Performance
|
| 45 |
+
|
| 46 |
|
| 47 |
## AGIEval Performance
|
| 48 |
+
|
| 49 |
|
| 50 |
We will report our results on [HuggingFaceH4 Open LLM Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard) Evals once we receive them.
|
| 51 |
|
|
|
|
| 77 |
year = {2023},
|
| 78 |
publisher = {HuggingFace},
|
| 79 |
journal = {HuggingFace repository},
|
| 80 |
+
howpublished = {\url{https://https://huggingface.co/Open-Orca/OpenOrca-Preview1-13B},
|
| 81 |
}
|
| 82 |
```
|
| 83 |
```bibtex
|