|
|
--- |
|
|
tags: |
|
|
- creativityneuro |
|
|
- llm-creativity |
|
|
- mechanistic-interpretability |
|
|
base_model: microsoft/Phi-3.5-mini-instruct |
|
|
license: apache-2.0 |
|
|
--- |
|
|
|
|
|
# phi-3.5-mini-instruct-cn-problem-kr0.2-a0.5-creative |
|
|
|
|
|
This is a **CreativityNeuro (CN)** modified version of [microsoft/Phi-3.5-mini-instruct](https://huggingface.co/microsoft/Phi-3.5-mini-instruct). |
|
|
|
|
|
## Model Details |
|
|
|
|
|
- **Base Model**: microsoft/Phi-3.5-mini-instruct |
|
|
- **Modification**: CreativityNeuro weight scaling |
|
|
- **Prompt Set**: problem |
|
|
- **Keep Ratio**: 0.2 (top 20.0% of task-specific weights) |
|
|
- **Alpha**: 0.5 (scaling strength) |
|
|
- **Mode**: creative |
|
|
|
|
|
## What is CreativityNeuro? |
|
|
|
|
|
CreativityNeuro identifies task-specific neurons using Wanda-style importance scoring and selectively |
|
|
upscales weights associated with creative thinking. The modification formula is: |
|
|
|
|
|
``` |
|
|
W_new = W × (1 + α × mask) |
|
|
``` |
|
|
|
|
|
Where `mask` identifies weights important for creative tasks but not for routine/associative tasks. |
|
|
|
|
|
## Usage |
|
|
|
|
|
```python |
|
|
from transformers import AutoModelForCausalLM, AutoTokenizer |
|
|
|
|
|
model = AutoModelForCausalLM.from_pretrained("priorcomputers/phi-3.5-mini-instruct-cn-problem-kr0.2-a0.5-creative") |
|
|
tokenizer = AutoTokenizer.from_pretrained("priorcomputers/phi-3.5-mini-instruct-cn-problem-kr0.2-a0.5-creative") |
|
|
|
|
|
# Use like any other model |
|
|
outputs = model.generate(...) |
|
|
``` |
|
|
|
|
|
## Citation |
|
|
|
|
|
If you use this model, please cite: |
|
|
|
|
|
```bibtex |
|
|
@misc{creativityneuro2025, |
|
|
title={CreativityNeuro: Mechanistic Interpretability for LLM Creativity}, |
|
|
author={Prior Computers}, |
|
|
year={2025}, |
|
|
url={https://huggingface.co/priorcomputers} |
|
|
} |
|
|
``` |
|
|
|
