Skip to content
View cjgwang's full-sized avatar
🌠
🌠
6 followers · 14 following
  • University of Oxford
  • Oxford, UK

Achievements

Achievement: Pull Sharkx2Achievement: Pair ExtraordinaireAchievement: YOLO

Achievements

Achievement: Pull Sharkx2Achievement: Pair ExtraordinaireAchievement: YOLO

Block or report cjgwang

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.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. 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
cjgwang/README.md

Hi!

I'm Cath Ge-Wang, an undergrad at Oxford studying Mathematics. I do technical AI safety research (AI alignment and AI control)

💻 Tech Stack

  • Languages: Python, JavaScript, C, R, SQL
  • ML Frameworks: Pytorch, scikit-learn
  • LLM Ecosystem: Hugging Face Transformers, TRL
  • Web: HTML, CSS, JavaScript, Flask
  • Tools & Platforms: Git, Docker, AWS, Inspect

🔎 Research

My primary research interests lie in AI control, evaluations, and model internals, particularly understanding and mitigating emergent misalignment risks in autonomous AI systems. I focus on empirical questions around goal misgeneralisation, alignment faking, and agentic evaluations, aiming to clarify failure modes in dangerous frontier models. I am also interested in how these technical insights inform AI governance and policy, especially hardware verification, mechanisms for strategic risk, and constraining dangerous capability deployment.

Last updated: Jun 2026

Pinned Loading

  1. tylercrosse/mars-attacks tylercrosse/mars-attacks Public

    Dynamic Attack Selection in Agentic AI Control Evaluations

    Python 1

  2. reward-multiplicity reward-multiplicity Public

    Reward multiplicity and using STARC metrics to train meaningfully diverse reward ensembles in a gridworld environment

    Python

  3. J0YY/patchsteg J0YY/patchsteg Public

    mech interp paper for oxford v. cambridge varsity hackathon (grand prize) - defeating SOTA steganographic attacks in latent space

    Python 1

  4. ARENA_3.0 ARENA_3.0 Public

    Forked from callummcdougall/ARENA_3.0

    Jupyter Notebook

  5. tea-brewer tea-brewer Public

    tea brewery website (flask)

    HTML

  6. sms-text-classifier sms-text-classifier Public

    NLP SMS classification using Naive Bayes

    Python