HaoYang Shang
I want to understand
why
things happen.
Email
CV
Scholar
GitHub
LinkedIn
mechanism
behavior
Projects
What Astra's weight edit actually programmed.
The model knew it was right. Thinking made it doubt.
Blog · 2026
Making human research reproducible.
HumanStudy Hub
How would we know an AI really predicted it?
Social Simulation Arena
Papers
Agents team up like brains, and overload like brains.
CoThinker
TMLR
A volume knob for cognitive bias in AI.
CoBRA
CHI 2026 · Best Paper
Memorization lives in a few tokens.
Mitigating Memorization Where It Happens
NeurIPS 2026
For fun
AI that goes on the date for you.
Love First, Know Later
NeurIPS Workshop · Oral
What does a simulation lose when it zooms out?
Measurable granularity for generative social simulation
Coming soon
