---
base_model:
- fblgit/cybertron-v4-qw7B-MGS
- huihui-ai/Qwen2.5-7B-Instruct-abliterated-v3
- FreedomIntelligence/HuatuoGPT-o1-7B
- rombodawg/Rombos-LLM-V2.5-Qwen-7b
- Qwen/Qwen2.5-7B-Instruct
library_name: transformers
tags:
- mergekit
- merge

---
# merge

This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).

## Merge Details
### Merge Method

This model was merged using the [Linear DELLA](https://arxiv.org/abs/2406.11617) merge method using [Qwen/Qwen2.5-7B-Instruct](https://huggingface.co/Qwen/Qwen2.5-7B-Instruct) as a base.

### Models Merged

The following models were included in the merge:
* [fblgit/cybertron-v4-qw7B-MGS](https://huggingface.co/fblgit/cybertron-v4-qw7B-MGS)
* [huihui-ai/Qwen2.5-7B-Instruct-abliterated-v3](https://huggingface.co/huihui-ai/Qwen2.5-7B-Instruct-abliterated-v3)
* [FreedomIntelligence/HuatuoGPT-o1-7B](https://huggingface.co/FreedomIntelligence/HuatuoGPT-o1-7B)
* [rombodawg/Rombos-LLM-V2.5-Qwen-7b](https://huggingface.co/rombodawg/Rombos-LLM-V2.5-Qwen-7b)

### Configuration

The following YAML configuration was used to produce this model:

```yaml
merge_method: della_linear
base_model: Qwen/Qwen2.5-7B-Instruct
dtype: bfloat16
parameters:
  epsilon: 0.015            # Fine-grain scaling for precision.
  lambda: 1.6               # Strong emphasis on top-performing models.
  normalize: true           # Stable parameter integration across models.
adaptive_merge_parameters:
  task_weights:
    tinyArc: 1.75           # Logical reasoning.
    tinyHellaswag: 1.65     # Contextual predictions.
    tinyMMLU: 1.8           # Domain knowledge.
    tinyTruthfulQA: 2.0     # Prioritize truthful reasoning.
    tinyTruthfulQA_mc1: 1.85
    tinyWinogrande: 1.9     # Advanced reasoning and predictions.
    IFEval: 2.1             # Instruction-following and multitasking.
    BBH: 2.25                # Complex reasoning.
    MATH: 2.4               # Mathematical reasoning.
    GPQA: 2.35             # Factual QA.
    MUSR: 2.3             # Multi-step reasoning.
    MMLU-PRO: 2.35       # Domain multitask performance.
  smoothing_factor: 0.05     # TURN UP THE SMOOTH!
models:
  - model: Qwen/Qwen2.5-7B-Instruct
    parameters:
      weight: 0.65
      density: 0.65
  - model: huihui-ai/Qwen2.5-7B-Instruct-abliterated-v3
    parameters:
      weight: 0.1
      density: 0.1
  - model: rombodawg/Rombos-LLM-V2.5-Qwen-7b
    parameters:
      weight: 0.15
      density: 0.15
  - model: fblgit/cybertron-v4-qw7B-MGS
    parameters:
      weight: 0.05
      density: 0.05
  - model: FreedomIntelligence/HuatuoGPT-o1-7B
    parameters:
      weight: 0.05
      density: 0.05

```