Understanding Self-Awareness in LLMs

commitment: 15 hrs/week
format: Research and writing
topic: Artificial sentience
open to mentee proposals: Yes, with the right mentee

mentored by

The project

Developing further behavior-based tests for endogenous goals in LLMs.
MechInterp investigations into the representations underlying apparent sandbagging and external cued increases in effort found in my recent research.
Investigating the causal role of chain-of-thought it apparently more self-aware behaviors in reasoning models.
Investigating the degree to which models maintain a persistent identity across contexts. One way to test this is to monitor pronoun usage, which has been linked to emerging self-awareness in children; when do models signal identification with whatever “part” they are playing with the user, vs their own identity as an AI?
Human studies: Establish a gold standard for self-awareness metrics to compare AIs against.
Conceptual: Build a better theoretical account of the components of self-awareness found in biology, and come up with other LLM-appropriate or architecture-agnostic paradigms to elicit self-awareness signatures.

The mentee's role

I want mentees who will be owners of a project to a degree that justifies first authorship on a paper. That can range from full implementation of a project agenda I specify in detail, to both implementing and substantially defining project direction.

Who I'm looking for

Comfortable writing substantial amounts of Python code, and facility with coding agents, for a technical project
High familiarity with LLMs as a user, and a solid understanding of how transformers work
Some background or at least interest in psychology/cognitive science and concepts of self and self-awareness
Familiarity with experimental design is a plus.

Questions for applicants

What are you talking to when you're talking to Claude/some other frontier-ish LLM?

Support offered

Shaping direction
Designing experiments
Domain expertise
Feedback
Writing

Christopher Ackerman

Independent

I'm an AI safety researcher and research manager interested in AI self-awareness and its implications. My empirical work focuses on objective, behavior-based evaluations of frontier LLMs.

I'm currently a Senior Research Manager at MATS Research and a Research Mentor with SPAR. Previously I was a Senior Quantitative UX Researcher at Google.
I hold a PhD in Neuroscience from Johns Hopkins, an MS in Computer Science from USC, and a BA in English Literature from the University of Chicago.