base_model: /Users/dawn/git/models/Starling-LM-7B-beta
gate_mode: hidden # one of "hidden", "cheap_embed", or "random"
dtype: bfloat16 # output dtype (float32, float16, or bfloat16)
experts:
- source_model: /Users/dawn/git/models/Silicon-Maid-7B
positive_prompts:
- "roleplay"
- source_model: /Users/dawn/git/models/Starling-LM-7B-beta
positive_prompts:
- "chat"
Open LLM Leaderboard Evaluation Results
Metric |
Value |
Avg. |
70.20 |
AI2 Reasoning Challenge (25-Shot) |
67.15 |
HellaSwag (10-Shot) |
85.00 |
MMLU (5-Shot) |
65.36 |
TruthfulQA (0-shot) |
57.98 |
Winogrande (5-shot) |
79.79 |
GSM8k (5-shot) |
65.88 |