eclectics — Moontae Lee on AI research and leadership

AI RESEARCH & LEADERSHIPMake hard problems measurable
Make hard problems measurable
Build teams that solve and ship
WHAT AI CAN DO
FOCUS / RESEARCH
DEER
What makes a research report worth trusting?
ICML 2026FIGURE FROM P.1
FOCUS / THOUGHTS
Double Helix of Intelligence (Part I)
The mathematics of keeping a generator and evaluator on course, from constrained initialization to game dynamics and mode selection.

RESEARCH ESSAYGENERATED FIGURE
RESEARCH, ON THE RECORD
SELECTED PAPERS
EMNLP 2026 FindingsC71
Gaming the Judge: How Unfaithful Chain-of-Thought Can Undermine Agent Evaluation
arXiv 2026P17
CompLit: Scientific Literature Search Benchmarks Must Cover Implicit, Cumulative, and Unmet Needs
NeurIPS 2025C58
MLRC-Bench: Can Language Agents Solve Machine Learning Research Challenges?
NeurIPS 2023C25
Projection Regret: Reducing Background Bias for Novelty Detection via Diffusion Models
ACL 2026C66
Adaptive Retrieval for Reasoning
NeurIPS Position Paper Track 2026C74
From Benchmark to Adoption: Coding Agent Evaluation Needs User-Level Harness
NAACL 2025C45
The BiGGen Bench: A Principled Benchmark for Fine-grained Evaluation of Language Models with Language Models
EMNLP 2024C43
Prometheus 2: An Open Source Language Model Specialized in Evaluating Other Language Models
EACL 2027C75
MERIT: From Preference Traces to Evaluation Standards for Community-Aligned LLM Evaluation
ACL 2026C67
Efficiently Learning To Reason or Not to Reason: Root-token Policy Optimization for Adaptive Thinking
CoLLA 2024C37
Towards More Diverse Evaluation of Class Incremental Learning: Representation Learning Perspective
arXiv 2026P13
Co-Evolving Actor-Conditioned Critics for Non-Verifiable Generation
ACL 2024 FindingsC39
Small Language Models Need Strong Verifiers to Self-Correct Reasoning
arXiv 2026P12
Financial Numerical Prediction and Allocation as Token Generation
ICML 2024C38
Degeneration-free Policy Optimization: RL Fine-Tuning for Language Models without Degeneration
NeurIPS 2026C73