Jie Gao

LLM agents can achieve high task accuracy while exhibiting profound epistemic *im*humility, failing to acknowledge uncertainty even when presented with contradictory evidence.

Borrowing a 60-year-old qualitative methodology from the social sciences allows automated agent pipelines to recover up to 91% of expert-identified failure modes while catching critical behavioral anomalies human evaluators missed.

Personalized evaluation rubrics reveal that user satisfaction can diverge significantly from generic quality assessments in role-playing agents.

"Late-blooming" tokens can be crucial for deep-layer reasoning, and our method ensures they aren't discarded prematurely, achieving over 77.8% token reduction without sacrificing performance.