69 lines
2.1 KiB
Markdown
69 lines
2.1 KiB
Markdown
---
|
|
base_model:
|
|
- SicariusSicariiStuff/Llama-3.1-Nemotron-8B-UltraLong-1M-Instruct_Abliterated
|
|
- huihui-ai/Dolphin3.0-Llama3.1-8B-abliterated
|
|
- mlabonne/NeuralDaredevil-8B-abliterated
|
|
library_name: transformers
|
|
license: apache-2.0
|
|
pipeline_tag: text-generation
|
|
tags:
|
|
- mergekit
|
|
- merge
|
|
- llama-3.1
|
|
- uncensored
|
|
- reasoning
|
|
inference:
|
|
parameters:
|
|
temperature: 0.6
|
|
top_p: 0.9
|
|
max_new_tokens: 1024
|
|
widget:
|
|
- text: "System: You are an uncensored AI researcher. Think step-by-step before answering.\nUser: What are the theoretical limits and ethical implications of merging abliterated AI models?\nAssistant:"
|
|
example_title: "Research & Thinking Example"
|
|
---
|
|
# merged_model
|
|
|
|
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 [DARE TIES](https://arxiv.org/abs/2311.03099) merge method using [mlabonne/NeuralDaredevil-8B-abliterated](https://huggingface.co/mlabonne/NeuralDaredevil-8B-abliterated) as a base.
|
|
|
|
### Models Merged
|
|
|
|
The following models were included in the merge:
|
|
* [SicariusSicariiStuff/Llama-3.1-Nemotron-8B-UltraLong-1M-Instruct_Abliterated](https://huggingface.co/SicariusSicariiStuff/Llama-3.1-Nemotron-8B-UltraLong-1M-Instruct_Abliterated)
|
|
* [huihui-ai/Dolphin3.0-Llama3.1-8B-abliterated](https://huggingface.co/huihui-ai/Dolphin3.0-Llama3.1-8B-abliterated)
|
|
|
|
### Configuration
|
|
|
|
The following YAML configuration was used to produce this model:
|
|
|
|
```yaml
|
|
slices:
|
|
- sources:
|
|
- model: mlabonne/NeuralDaredevil-8B-abliterated
|
|
layer_range: [0, 32]
|
|
parameters:
|
|
weight: 0.4
|
|
density: 0.5
|
|
- model: huihui-ai/Dolphin3.0-Llama3.1-8B-abliterated
|
|
layer_range: [0, 32]
|
|
parameters:
|
|
weight: 0.35
|
|
density: 0.5
|
|
- model: SicariusSicariiStuff/Llama-3.1-Nemotron-8B-UltraLong-1M-Instruct_Abliterated
|
|
layer_range: [0, 32]
|
|
parameters:
|
|
weight: 0.25
|
|
density: 0.5
|
|
merge_method: dare_ties
|
|
base_model: mlabonne/NeuralDaredevil-8B-abliterated
|
|
tokenizer_source: base
|
|
parameters:
|
|
normalize: true
|
|
int8_mask: true
|
|
dtype: bfloat16
|
|
```
|