Skip to content
View mash-py's full-sized avatar
🏠
Working from home
🏠
Working from home

Block or report mash-py

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
mash-py/README.md
Hello Earth! typing 010101

she/her · python teacher · AI researcher · curious developer · vegan · traveler

010101

🍄 about me

class AboutMe:
    name       = "Maria"
    role       = ["Python Teacher 🐍", "AI Researcher 🔬", "Developer 💻"]
    based_in   = "🌍 Earth (nomad since 2023)"
    languages  = ["Python", "English", "Russian"]
    hobbies    = ["coding", "teaching", "Dota 2", "anime", "learning", "travel"]
    currently  = "researching how neural networks think from the inside"
    learning   = ["C", "ML", "Math", "Computer Science", "System Design"]
    belief     = "Consciousness is like a protocol"

🐸 what i'm working on

🔬 Interp Research

Independent mechanistic interpretability research — looking for directions inside models that shape their behavior.

🐍 PYTHONCRAFT

My own Python course for kids — from first print() to games in Pygame.

🎨 WEBCRAFT

An HTML/CSS course for kids — building first websites step by step.


🌿 skills & tools

languages

Python HTML5 CSS3

frameworks & libraries

Django Pygame NumPy

ML & interpretability

PyTorch Hugging Face einops Ollama

tools

Git GitHub VS Code Claude Code


🐌 find me

Portfolio Telegram Instagram


010101 be happy, stay curious and learn more

Popular repositories Loading

  1. shutdown-self-relevance-llms shutdown-self-relevance-llms Public

    Difference-of-means "shutdown directions" in the residual stream of 10 LLMs (Qwen2.5, Llama, Mistral, Phi, Gemma). A placebo control shows arbitrary neutral prompt pairs separate just as well, so n…

    Python 1

  2. llm-food-animal-axis llm-food-animal-axis Public

    LLMs encode a linear "food animal" direction: it makes harming farmed animals seem more acceptable without lowering their perceived capacity to suffer. Preregistered study on Qwen2.5-7B, Llama-3.1…

    Python 1

  3. mash-py mash-py Public

    Mash

  4. webcraft webcraft Public

    Курс по html/css для детей от 7 лет!

  5. Mash Mash Public

    my personal website

    HTML