mayacinka/ramonda-7b-dpo-ties
mayacinka/ramonda-7b-dpo-ties is a 7 billion parameter language model created by mayacinka, formed by merging paulml/OGNO-7B and bardsai/jaskier-7b-dpo-v4.3 using the TIES merging method. This model achieves an average score of 76.19 on the Open LLM Leaderboard and 62.12 on LLM AutoEval, demonstrating strong performance across various reasoning and language understanding tasks. It is suitable for general-purpose applications requiring robust language generation and comprehension within a 4096-token context window.
Loading preview...
Model Overview
mayacinka/ramonda-7b-dpo-ties is a 7 billion parameter language model developed by mayacinka. It is a merged model, combining the strengths of paulml/OGNO-7B and bardsai/jaskier-7b-dpo-v4.3 using the TIES merging method. This approach aims to leverage the distinct capabilities of its constituent models to achieve enhanced performance.
Key Performance Metrics
The model demonstrates competitive results on standard benchmarks:
- Open LLM Leaderboard Average: 76.19
- ARC: 72.7
- HellaSwag: 89.69
- MMLU: 64.5
- TruthfulQA: 77.17
- Winogrande: 84.77
- GSM8K: 68.92
- LLM AutoEval Average: 62.12
- AGIEval: 44.67
- GPT4All: 77.16
- TruthfulQA: 77.6
- Bigbench: 49.06
Usage and Configuration
The model is designed for text generation tasks and can be easily integrated into Python environments using the transformers library. Its configuration specifies a ties merge method with bardsai/jaskier-7b-dpo-v5.6 as the base model, and specific density and weight parameters for the merged components. The model supports float16 dtype for efficient inference.