Benyou Wang

Human or Machine? A Preliminary Turing Test for Speech-to-Speech Interaction

Lang-Prune: Unlocking Fair and Powerful Pruning for Multilingual Large Language Models

EchoMind: An Interrelated Multi-level Benchmark for Evaluating Empathetic Speech Language Models

EvA: An Evidence-First Audio Understanding Paradigm for LALMs

Does Higher Interpretability Imply Better Utility? A Pairwise Analysis on Sparse Autoencoders

Towards Fine-grained Audio Captioning with Multimodal Contextual Fusion

LiveClin: A Live Clinical Benchmark without Leakage

On the Scaling Flaws of Verifier-Guided Beam Search

TTS-Hub: Leveraging Modular LoRAs and Arithmetic Composition for Controllable Text-to-Speech

MicroVerse: A Preliminary Exploration Toward a Micro-World Simulation

AV-Odyssey Bench: From Fundamental Audio Perception to Audio-Visual Understanding

CALM Before the STORM: Unlocking Native Reasoning for Optimization Modeling

Benchmarking LLMs on Authentic Cases from Medical Journals

Learning from Peers in Reasoning Models

DualRes: A Resampling-Based Framework for Enhancing Probabilistic Forecasting

EchoX: Towards Mitigating Acoustic-Semantic Gap via Echo Training for Speech-to-Speech LLMs

Uncertainty-Aware Search and Value Models in LLMs

Video-R1: Reinforcing Video Reasoning in MLLMs

QFFT, Question-Free Fine-Tuning for Adaptive Reasoning

TwinMarket: A Scalable Behavioral and Social Simulation for Financial Markets

Omni-MATH: A Universal Olympiad Level Mathematic Benchmark for Large Language Models

Efficiently Democratizing Medical LLMs for 50 Languages via a Mixture of Language Family Experts

Teaching Language Models to Reason with Tools

Self-Evolving Critique Abilities in Large Language Models

The First Few Tokens Are All You Need: An Efficient and Effective Unsupervised Prefix Fine-Tuning Method for Reasoning Models

E$^2$-RAG: Towards Editable Efficient RAG by Editing Compressed KV Caches

LongLLaVA: Scaling Multi-modal LLMs to 1000 Images Efficiently via a Hybrid Architecture

Towards Understanding Fine-Tuning Mechanisms of LLMs via Circuit Analysis

Synthetic Data (Almost) from Scratch: Generalized Instruction Tuning for Language Models

Vector Quantization By Distribution Matching

Periodical Moving Average Accelerates Gradient Accumulation for Post-Training

Roadmap towards Superhuman Speech Understanding using Large Language Models