google-deepmind/gemma_penzai
A JAX Research Toolkit for Visualizing, Manipulating, and Understanding Gemma Models with Multi-modal Support based on Penzai.
This toolkit helps AI researchers and interpretability scientists explore and understand how multimodal Large Language Models (LLMs) like Gemma 3 process information. It takes an existing Gemma model (including those with vision capabilities) and allows you to visualize and manipulate its internal mechanisms. The output helps researchers gain deeper insights into model behavior.
Use this if you are an AI researcher or safety scientist who needs to perform mechanistic interpretability on Gemma models, especially multimodal versions, to understand their internal workings.
Not ideal if you are a developer looking to simply deploy or fine-tune Gemma models without needing to deeply analyze their internal computational graphs and behaviors.
Stars
90
Forks
5
Language
Jupyter Notebook
License
Apache-2.0
Category
Last pushed
Jan 13, 2026
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/transformers/google-deepmind/gemma_penzai"
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