4 papers
Prompt Injection as Role Confusion
Charles Ye, Jasmine Cui, Dylan Hadfield-Menell
LLMs see the world as a single stream of text, partitioned into roles like <user> or <tool>. We trace prompt injection to role confusion: models perceive the source of text from ho…
Polysemantic Experts, Monosemantic Paths: Routing as Control in MoEs
Charles Ye, Bo Yuan, Lee Sharkey
An LLM's residual stream is both state and instruction: it encodes the current context and determines the next transformation. We introduce a parameter-free decomposition for Mixtu…
Emergent Search and Backtracking in Latent Reasoning Models
Jasmine Cui, Charles Ye
What happens when a language model thinks without words? Standard reasoning LLMs verbalize intermediate steps as chain-of-thought; latent reasoning transformers (LRTs) instead perf…
Efficient Representations are Controllable Representations
Charles Ye, Jasmine Cui
What is the most brute-force way to install interpretable, controllable features into a model's activations? Controlling how LLMs internally represent concepts typically requires s…