LESSWRONG
LW

Ruixuan Huang

11210

An undergraduate student from USTC CSE. Currently during internship in MSRA social computing group. Recent research interests are XAI-related issues.

Posts

Sorted by New

Wikitag Contributions

Comments

Sorted by

Subspace Rerouting: Using Mechanistic Interpretability to Craft Adversarial Attacks against Large Language Models

Ruixuan Huang5mo10

Great job! Consider reading our related paper: https://arxiv.org/abs/2404.12038

No wikitag contributions to display.

8Steering LLMs' Behavior with Concept Activation Vectors

1y

0

6Exploring the Evolution and Migration of Different Layer Embedding in LLMs

1y

0