Yejin Choi

In-the-Flow Agentic System Optimization for Effective Planning and Tool Use

Unfolding Spatial Cognition: Evaluating Multimodal Models on Visual Simulations

ProfBench: Multi-Domain Rubrics requiring Professional Knowledge to Answer and Judge

OpenThoughts: Data Recipes for Reasoning Models

LitmusValues: Will AI Tell Lies to Save Sick Children? Litmus-Testing AI Values Prioritization with AIRiskDilemmas

Theory of Space: Can Foundation Models Construct Spatial Beliefs through Active Exploration?

Multiplayer Nash Preference Optimization

MoReBench: Evaluating Procedural and Pluralistic Moral Reasoning in Language Models, More than Outcomes

RLP: Reinforcement as a Pretraining Objective

Learning Human-Perceived Fakeness in AI-Generated Videos via Multimodal LLMs

Goedel-Prover-V2: Scaling Formal Theorem Proving with Scaffolded Data Synthesis and Self-Correction

Chasing Moving Targets with Online Self-Play Reinforcement Learning for Safer Language Models

ThinkMorph: Emergent Properties in Multimodal Interleaved Chain-of-Thought Reasoning

Front-Loading Reasoning: The Synergy between Pretraining and Post-Training Data

The Invisible Leash? Why RLVR May or May Not Escape Its Origin

Spectrum Tuning: Post-Training for Distributional Coverage and In-Context Steerability

ODESteer: A Unified ODE-Based Steering Framework for LLM Alignment

DLER: Doing Length pEnalty Right — Incentivizing More Intelligence per Token via Reinforcement Learning

PoseX: AI Defeats Physics-based Methods on Protein Ligand Cross-Docking

ToolOrchestra: Elevating Intelligence via Efficient Model and Tool Orchestration

HieraSuite: A Holistic Toolkit for Building Versatile System-User Instruction Hierarchy

DeepSearch: Overcome the Bottleneck of Reinforcement Learning with Verifiable Rewards via Tree-based Search

SimpleToM: Exposing the Gap between Explicit ToM Inference and Implicit ToM Application in LLMs

RiboPO: Preference Optimization for Structure- and Stability-Aware RNA Design

L^2M^3OF: A Large Language Multimodal Model for Metal-Organic Frameworks

Privacy-Preserving LLM Interaction with Socratic Chain-of-Thought Reasoning and Homomorphically Encrypted Vector Databases

BroRL: Scaling Reinforcement Learning via Broadened Exploration

Long Grounded Thoughts: Distilling Compositional Visual Reasoning Chains at Scale

Privasis: Synthesizing the Largest "Public" Private Dataset from Scratch

Position: The Hidden Costs and Measurement Gaps of Reinforcement Learning with Verifiable Rewards

TinyEye: Sharpening Visual Reasoning of Tiny Models with Offline Policy Optimization

Are you going to finish that? A Practical Study of the Tokenization Boundary Problem

When Visualizing is the First Step to Reasoning: MIRA, a Benchmark for Visual Chain-of-Thought

UQ: Assessing Language Models on Unsolved Questions

Fighting Fire with Fire: Assessing Test Set Contamination Through Deliberate Training on Test Data

iGRPO: Self‑Feedback–Driven LLM Reasoning

The Agent's Marathon: Probing the Limits of Endurance in Long-Horizon Tasks

From Supervision to Exploration: What Does Protein Language Model Learn During Reinforcement Learning?

Prismatic Synthesis: Gradient-based Data Diversification Boosts Generalization in LLM Reasoning

Trust or Escalate: LLM Judges with Provable Guarantees for Human Agreement

SuperBPE: Space Travel for Language Models

Why and How LLMs Hallucinate: Connecting the Dots with Subsequence Associations

Broken Tokens? Your Language Model can Secretly Handle Non-Canonical Tokenizations

Model Swarms: Collaborative Search to Adapt LLM Experts via Swarm Intelligence

WildBench: Benchmarking LLMs with Challenging Tasks from Real Users in the Wild

ProRL: Prolonged Reinforcement Learning Expands Reasoning Boundaries in Large Language Models

VAGEN: Reinforcing World Model Reasoning for Multi-Turn VLM Agents

DailyDilemmas: Revealing Value Preferences of LLMs with Quandaries of Daily Life

Language Model Alignment in Multilingual Trolley Problems

AI as Humanity’s Salieri: Quantifying Linguistic Creativity of Language Models via Systematic Attribution of Machine Text against Web Text

Verifying the Verifiers: Unveiling Pitfalls and Potentials in Fact Verifiers

Model Swarms: Collaborative Search to Adapt LLM Experts via Swarm Intelligence

HAICOSYSTEM: An Ecosystem for Sandboxing Safety Risks in Interactive AI Agents

AI Debate Aids Assessment of Controversial Claims

The Surprising Effectiveness of Membership Inference with Simple N-Gram Coverage

Hypothesis-Driven Theory-of-Mind Reasoning for Large Language Models

HAICOSYSTEM: An Ecosystem for Sandboxing Safety Risks in Human-AI Interactions

Explore Theory of Mind: program-guided adversarial data generation for theory of mind reasoning

CertainlyUncertain: A Benchmark and Metric for Multimodal Epistemic and Aleatoric Awareness

X-Teaming: Multi-Turn Jailbreaks and Defenses with Adaptive Multi-Agents

A False Sense of Privacy: Evaluating Textual Data Sanitization Beyond Surface-level Privacy Leakage

Magpie: Alignment Data Synthesis from Scratch by Prompting Aligned LLMs with Nothing

Diverging Preferences: When do Annotators Disagree and do Models Know?

Diverging Preferences: When do Annotators Disagree and do Models Know?

ZebraLogic: On the Scaling Limits of LLMs for Logical Reasoning

SafetyAnalyst: Interpretable, Transparent, and Steerable Safety Moderation for AI Behavior

Benchmarking Vision Language Model Unlearning via Fictitious Facial Identity Dataset

SimpleToM: Exposing the Gap between Explicit ToM Inference and Implicit ToM Application in LLMs

LongPerceptualThoughts: Distilling System-2 Reasoning for System-1 Perception

CulturalBench: a Robust, Diverse and Challenging Benchmark on Measuring (the Lack of) Cultural Knowledge of LLMs

From Explicit CoT to Implicit CoT: Learning to Internalize CoT Step by Step

Can Language Models Reason about Individualistic Human Values and Preferences?

Pixelated Instructions: Can Multimodal Large Language Models Follow Printed Instructions in Images?

SafetyAnalyst: Interpretable, transparent, and steerable LLM safety moderation

The HALoGen Benchmark: Fantastic LLM Hallucinations and Where To Find Them