Update README.md
Browse files
README.md
CHANGED
|
@@ -38,6 +38,24 @@ The following models were included in the merge:
|
|
| 38 |
>A mixture of [Sao10K/Solstice-11B-v1](https://huggingface.co/Sao10K/Solstice-11B-v1) and
|
| 39 |
>[ABX-AI/Fimbulvetr-Kuro-Lotus-v2-10.7B] which is updated saishf/Fimbulvetr-Kuro-Lotus-10.7B with Fimb v2
|
| 40 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 41 |
### Configuration
|
| 42 |
|
| 43 |
The following YAML configuration was used to produce this model:
|
|
|
|
| 38 |
>A mixture of [Sao10K/Solstice-11B-v1](https://huggingface.co/Sao10K/Solstice-11B-v1) and
|
| 39 |
>[ABX-AI/Fimbulvetr-Kuro-Lotus-v2-10.7B] which is updated saishf/Fimbulvetr-Kuro-Lotus-10.7B with Fimb v2
|
| 40 |
|
| 41 |
+
### OpenLLM Eval Results
|
| 42 |
+
|
| 43 |
+
[Detailed Results + Failed GSM8K](https://huggingface.co/datasets/open-llm-leaderboard/details_ABX-AI__Silver-Sun-v2-11B)
|
| 44 |
+
|
| 45 |
+
|
| 46 |
+
>[NOTE] I had to remove GSM8K from the results and manually re-average the rest. GSM8K failed inexplicably, and it should not have.
|
| 47 |
+
>By removing the GSM8K score, the average is VERY close to upstage/SOLAR-10.7B-v1.0 (74.20), which would make sense.
|
| 48 |
+
|
| 49 |
+
| Metric |Value|
|
| 50 |
+
|---------------------------------|----:|
|
| 51 |
+
|Avg. |74.04|
|
| 52 |
+
|AI2 Reasoning Challenge (25-Shot)|69.88|
|
| 53 |
+
|HellaSwag (10-Shot) |87.81|
|
| 54 |
+
|MMLU (5-Shot) |66.74|
|
| 55 |
+
|TruthfulQA (0-shot) |62.49|
|
| 56 |
+
|Winogrande (5-shot) |83.27|
|
| 57 |
+
|
| 58 |
+
|
| 59 |
### Configuration
|
| 60 |
|
| 61 |
The following YAML configuration was used to produce this model:
|