Skip to content
View sagnikc395's full-sized avatar

Block or report sagnikc395

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. 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
sagnikc395/README.md

hi, i'm sagnik

i'm passionate about AI Alignment and post-training techniques for LLMs
currently focussing on mechanistic interpretability and AI Control.

some writeups of thoughts and projects

Pinned Loading

  1. refusal-steering-using-saes refusal-steering-using-saes Public

    Do SAE Features Outperform Simple Baselines for Detecting and Steering Refusal? [NEMI Workshop 2026]

    Jupyter Notebook

  2. crisp-reproducibility crisp-reproducibility Public

    Reproducibility of CRISP: Persistent Concept Unlearning via Sparse Autoencoders[BluedotAI Technical AI Safety Project]

    Jupyter Notebook

  3. anchora anchora Public

    an workflow engine for agents

    Go

  4. kai kai Public

    a agent harness

    Python

  5. refusal-circuit-dpo refusal-circuit-dpo Public

    in a tiny-parameter base model, which layers , attention heads, and residual stream directions are responsible for the emergent refusal behavior introduced by DPO

    Python

  6. obsfucation-tax obsfucation-tax Public

    what does evading an activation probe actually cost an attacker inside a control protocol ?

    Python