AI Research
We study Creative Quality Alignment (CQA) — using information theory to formalize what makes creative work "good" and to align AI systems accordingly. Our research program comprises a 3+1 paper series: one theoretical foundation, one experimental validation, one evaluation benchmark, and one expert-knowledge extraction paradigm.
Calibrated Surprise: An Information-Theoretic Account of Creative Quality
Preprint
Uses Shannon's mutual information to prove that full-dimensional accuracy and mediocrity are mutually exclusive.
[arXiv]
Creative Quality Alignment: Expert Tacit Knowledge Transfer via Chain-of-Thought Fine-Tuning
In Preparation
Empirical validation of the Calibrated Surprise framework via LLM fine-tuning experiments.
QUIET: Quality Understanding via Implicit Evaluation Testing
In Preparation
A professional evaluation benchmark for creative quality alignment.
BC Protocol: Structured Dual-Expert Dialogue for Eliciting High-Quality Chain-of-Thought Post-Training Data
In Preparation
A structured dialogue paradigm for eliciting latent expert reasoning chains.
AI Alignment
Creative Quality
Information Theory
Chain-of-Thought
Expert Knowledge Extraction
