Yulia Tsvetkov

Don't Throw Away Your Pretrained Model

MMMG: A Comprehensive and Reliable Benchmark for Multitask Multimodal Generation

$MADFormer$: Mixed Autoregressive and Diffusion Transformers for Continuous Image Generation

GuessBench: Sensemaking Multimodal Creativity in the Wild

PrefDisco: Benchmarking Proactive Personalized Reasoning

BLAB: Brutally Long Audio Bench

Spurious Rewards: Rethinking Training Signals in RLVR

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

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

MARVEL: Modular Abstention for Reliable and Versatile Expert LLMs

Generalizable LLM Learning of Graph Synthetic Data with Post-training Alignment

Irrelevant Context Helps: Understanding the Impact of Context in Large Language Models

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

Sparta Alignment: Collectively Aligning Multiple Language Models through Combat

ALFA: Aligning LLMs to Ask Good Questions A Case Study in Clinical Reasoning

Heterogeneous Swarms: Jointly Optimizing Model Roles and Weights for Multi-LLM Systems

Precise Information Control in Long-Form Text Generation

PrefPalette: Personalized Preference Modeling with Latent Attributes

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

Varying Shades of Wrong: Aligning LLMs with Wrong Answers Only

Finding Flawed Fictions: Evaluating Complex Reasoning in Language Models via Plot Hole Detection

JPEG-LM: LLMs as Image Generators with Canonical Codec Representations

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

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

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

Conformal Reasoning: Uncertainty Estimation in Interactive Environments