README.md

metadata

license: apache-2.0

---

    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