Update README.md
Browse files
README.md
CHANGED
|
@@ -49,6 +49,23 @@ The following models were included in the merge:
|
|
| 49 |
>A mixture of [Sao10K/Solstice-11B-v1](https://huggingface.co/Sao10K/Solstice-11B-v1) and
|
| 50 |
>[ABX-AI/Fimbulvetr-Kuro-Lotus-v2-10.7B] which is updated saishf/Fimbulvetr-Kuro-Lotus-10.7B with Fimb v2
|
| 51 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 52 |
### Configuration
|
| 53 |
|
| 54 |
The following YAML configuration was used to produce this model:
|
|
|
|
| 49 |
>A mixture of [Sao10K/Solstice-11B-v1](https://huggingface.co/Sao10K/Solstice-11B-v1) and
|
| 50 |
>[ABX-AI/Fimbulvetr-Kuro-Lotus-v2-10.7B] which is updated saishf/Fimbulvetr-Kuro-Lotus-10.7B with Fimb v2
|
| 51 |
|
| 52 |
+
### OpenLLM Eval Results
|
| 53 |
+
|
| 54 |
+
[Detailed Results + Failed GSM8K](https://huggingface.co/datasets/open-llm-leaderboard/details_ABX-AI__Silver-Sun-v2-11B)
|
| 55 |
+
|
| 56 |
+
|
| 57 |
+
>[NOTE] I had to remove GSM8K from the results and manually re-average the rest. GSM8K failed inexplicably, and it should not have.
|
| 58 |
+
>By removing the GSM8K score, the average is VERY close to upstage/SOLAR-10.7B-v1.0 (74.20), which would make sense.
|
| 59 |
+
|
| 60 |
+
| Metric |Value|
|
| 61 |
+
|---------------------------------|----:|
|
| 62 |
+
|Avg. |74.04|
|
| 63 |
+
|AI2 Reasoning Challenge (25-Shot)|69.88|
|
| 64 |
+
|HellaSwag (10-Shot) |87.81|
|
| 65 |
+
|MMLU (5-Shot) |66.74|
|
| 66 |
+
|TruthfulQA (0-shot) |62.49|
|
| 67 |
+
|Winogrande (5-shot) |83.27|
|
| 68 |
+
|
| 69 |
### Configuration
|
| 70 |
|
| 71 |
The following YAML configuration was used to produce this model:
|