babycommando/neuralgraffiti
Live-bending a foundation model’s output at neural network level.
This project helps anyone working with large language models to subtly influence their behavior and 'personality' over time without needing to retrain them. It takes a pre-trained LLM and a series of interactions or inputs, then, instead of specific words, it gradually shifts the model's internal 'thinking' to reflect certain concepts or leanings. The result is an LLM that develops a unique, consistent tone or conceptual preference over many interactions, like a digital persona.
273 stars. No commits in the last 6 months.
Use this if you want an existing large language model to develop a consistent 'memory' or 'personality' that subtly influences its responses over time, rather than just generating a one-off answer.
Not ideal if you need a model to consistently produce exact phrases or follow strict factual guidelines, as this tool influences general conceptual leanings rather than specific outputs.
Stars
273
Forks
24
Language
Jupyter Notebook
License
—
Category
Last pushed
Apr 07, 2025
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/transformers/babycommando/neuralgraffiti"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
unslothai/unsloth
Fine-tuning & Reinforcement Learning for LLMs. 🦥 Train OpenAI gpt-oss, DeepSeek, Qwen, Llama,...
huggingface/peft
🤗 PEFT: State-of-the-art Parameter-Efficient Fine-Tuning.
modelscope/ms-swift
Use PEFT or Full-parameter to CPT/SFT/DPO/GRPO 600+ LLMs (Qwen3.5, DeepSeek-R1, GLM-5,...
oumi-ai/oumi
Easily fine-tune, evaluate and deploy gpt-oss, Qwen3, DeepSeek-R1, or any open source LLM / VLM!
linkedin/Liger-Kernel
Efficient Triton Kernels for LLM Training