Model Overview
AI-B/UTENA-7B-NSFW-V2 is a 7 billion parameter language model developed by AI-B, created through a strategic merge of two base models: AI-B/UTENA-7B-NSFW and AI-B/UTENA-7B-BAGEL. This merge was performed using the SLERP (Spherical Linear Interpolation) method, a technique known for combining the strengths of different models while maintaining coherence.
Key Capabilities & Performance
This model is specifically engineered for particular content generation tasks, building upon its merged components. It demonstrates a balanced performance across various benchmarks, as indicated by its evaluation on the Open LLM Leaderboard:
- Average Score: 63.45
- AI2 Reasoning Challenge (25-Shot): 63.31
- HellaSwag (10-Shot): 84.54
- MMLU (5-Shot): 63.97
- TruthfulQA (0-shot): 47.81
- Winogrande (5-shot): 78.69
- GSM8k (5-shot): 42.38
These scores highlight its proficiency in reasoning, common sense, and factual question answering, alongside mathematical problem-solving. The model operates with a context length of 4096 tokens.
Usage Considerations
Users should note that this model is a merge of specific base models, suggesting its design is tailored for particular applications. Quantized versions, such as UTENA-7B-NSFW-V2-GGUF, are available for optimized deployment.