Preprints
SPADE: Self-Play in Adaptive Synthetic Executable Environments
Demystifying Reinforcement Learning Post-Training of Language Models
Forty Shades of Blue: Quality-Diversity Alignment via Mode-Conditioned Reinforcement Learning
How LLMs Distort Our Written Language
Tapes Together Strong: The Co-evolution of Computation and Cooperation
AI Epistemic Risks: Emerging Mechanisms & Evidence
Learn to Match: Two-Sided Matching with Temporally Extended Feedback
Beyond Cooperative Simulators: Generating Realistic User Personas for Robust Evaluation of LLM Agents
COLM 2026 Workshop on Social Simulation with LLMs
Evaluating & Reducing Deceptive Dialogue From Language Models with Multi-turn RL
AgenticRed: Evolving Agentic Systems for Red-Teaming
Are Language Models Sensitive to Morally Irrelevant Distractors?
Do as the Romans Do: Learning Universal Behaviors from Heterogeneous Agents
Embedded-Arena: Building Hardware-in-the-Loop Coding Agents to Run AI on Microcontrollers
FlowSteer: Prompt-Only Workflow Steering Exposes Planning-Time Vulnerabilities in Multi-Agent LLM Systems
2026
Chasing Moving Targets with Online Self-Play Reinforcement Learning for Safer Language Models
International Conference on Machine Learning (ICML) 2026
(Outstanding Paper Award, Oral Presentation) @ COLM 2025 AIA Workshop
(Outstanding Paper Award, Oral Presentation) @ COLM 2025 AIA Workshop
RLVE: Scaling Up Reinforcement Learning for Language Models with Adaptive Verifiable Environments
International Conference on Machine Learning (ICML) 2026
Learning to Summarize User Information for Personalized Reinforcement Learning from Human Feedback
International Conference on Learning Representations (ICLR) 2026
Generative Adversarial Post-Training Mitigates Reward Hacking in Live Human-AI Music Interaction
International Conference on Learning Representations (ICLR) 2026
AutoCode: LLMs as Problem Setters for Competitive Programming
International Conference on Learning Representations (ICLR) 2026
Improving Human-AI Coordination through Online Adversarial Training and Generative Models
International Conference on Learning Representations (ICLR) 2026
SPIRAL: Self-Play on Zero-Sum Games Incentivizes Reasoning via Multi-Agent Multi-Turn Reinforcement Learning
International Conference on Learning Representations (ICLR) 2026
Modeling Others' Minds as Code
International Conference on Learning Representations (ICLR) 2026 (Best Paper Award, Oral Presentation) @ NeurIPS 2025 LAW Workshop
Improving Interactive In-Context Learning from Natural Language Feedback
Google DeepMind Technical Report 2026
Pigeonholing: How Bad Prompts Hurt Models, Causing Collapse and Mistakes
Empirical Methods in Natural Language Processing (EMNLP) 2026 Findings
Maximizing Mutual Information Between Prompt and Response Improves LLM Performance with No Additional Data
International Conference on Machine Learning (ICML) 2026
LongCoT: Benchmarking Long-Horizon Chain-of-Thought Reasoning
International Conference on Machine Learning (ICML) 2026
Position: Solipsistic Superintelligence Is Unlikely to Be Cooperative
International Conference on Machine Learning (ICML) 2026 (Position Track)
2025
Enhancing Personalized Multi-Turn Dialogue with Curiosity Reward
Neural Information Processing Systems (NeurIPS) 2025
Madrona Prize @ Paul G. Allen School of Computer Science & Engineering
Madrona Prize @ Paul G. Allen School of Computer Science & Engineering
Consistently Simulating Human Personas with Multi-Turn Reinforcement Learning
Neural Information Processing Systems (NeurIPS) 2025
Evaluating Generalization Capabilities of LLM-Based Agents in Mixed-Motive Scenarios Using Concordia
Neural Information Processing Systems (NeurIPS) 2025
Achieving Human Level Competitive Robot Table Tennis
IEEE International Conference on Robotics and Automation (ICRA) 2025 (Best Paper Finalist)
Cross-environment Cooperation Enables Zero-shot Multi-agent Coordination
International Conference on Machine Learning (ICML) 2025 (Oral Paper-Top 1%) and CogSci 2025
Multi Agent Reinforcement Learning for Sequential Satellite Assignment Problems
AAAI Conference on Artificial Intelligence (AAAI) 2025 (Oral Paper - Top 5%)
Infer Human’s Intentions Before Following Natural Language Instructions
AAAI Conference on Artificial Intelligence (AAAI) 2025
ReaLJam: Real-Time, Synchronous Human-AI Music Jamming with Reinforcement Learning-Tuned Transformers
Extended Abstracts of The ACM Conference on Human Factors in Computing Systems (CHI) 2025
An Efficient Open World Benchmark for Multi-Agent Reinforcement Learning
NeurIPS Open World Agents Workshop 2025
Generative Modeling for Robust Deep Reinforcement Learning on the Traveling Salesman Problem
NeurIPS MATH.AI Workshop 2025
InvestESG: A multi-agent reinforcement learning benchmark for studying climate investment as a social dilemma
International Conference on Learning Representations (ICLR) 2025
2024
Personalizing Reinforcement Learning from Human Feedback with Variational Preference Learning
Neural Information Processing Systems (NeurIPS) 2024 (Spotlight-Top 2%)
Learning to Cooperate with Humans Using Generative Agents
Neural Information Processing Systems (NeurIPS) 2024
Impossibility theorems for feature attribution
Proceedings of the National Academy of Sciences (PNAS) 2024
The Concordia Contest: Advancing the Cooperative Intelligence of Language Agents
Neural Information Processing Systems (NeurIPS) Competition Track 2024
Moral Foundations of Large Language Models
Empirical Methods in Natural Language Processing (EMNLP) 2024 (Best Paper, AAAI Workshop on Representation Learning for Responsible Human-Centric AI)
Adaptive Accompaniment with ReaLchords
International Conference on Machine Learning (ICML), 2024