Noushin Quazi
United States
PhD Student
Carnegie Mellon University (CMU)
Neuroscience/Cognitive Science
United States PhD Student Carnegie Mellon University (CMU) Neuroscience/Cognitive Science
Project
Testing and Stress-Testing Claims of Global Workspace, World Models, and Valence Dynamics in Large Language Model Agents
Project Overview
There are widespread claims that contemporary large language models (LLMs) exhibit properties often associated with cognition in biological systems, such as global workspace–like coordination, predictive world models, and emotion- or valence-like dynamics. However, many of these claims rely on loose analogies, post-hoc interpretations, or descriptive probes, and it remains unclear which proposed signatures are empirically robust, well-defined, or causally relevant to behavior.
This project adopts a deliberately critical and falsification-oriented approach. Rather than asking whether LLM/VLA agents “have” these properties, we will design empirical tests and interventions that attempt to break commonly cited indicators. The aim is to clarify which claims withstand careful scrutiny, which collapse under intervention, and where current models systematically fail. This aligns with the program’s goal of critically examining sentience-like features without presupposing any particular stance on AI sentience.
Mentors
About the Scholar
My interests lie at the intersection of world models, reinforcement learning, and neuroethology. I am particularly interested in understanding the relationship between brain and behavior through probabilistic inference and model-based reinforcement learning, and how resulting insights can help develop algorithms for decision-making in partially observable, non-stationary environments. Towards these goals, I have been modeling the behavior of freely moving animals doing partially observable tasks using the POMDP framework. In my free time, I like to read and write poetry, go to parks, and watch sci-fi horror flicks.
Why AISS?
"It is inevitable that AI will emerge with capabilities once held sacred to humanity. I want to be a part of a community that recognizes this and espouses one of humanity’s greatest strengths when it comes to how we deal with sentient AI: empathy and compassion."