Kun Zhou

Exploring Expert Concentration for Parameter-efficient Fine-tuning of Mixture-of-Expert LLMs

Unleashing Perception-Time Scaling to Multimodal Reasoning Models

Codified Finite-state Machines for Role-playing

Model-aware Counterfactual Data based Contrastive Decoding for Video-LLM

Learning Modal-mixed Chain-of-thought Reasoning with Latent Embedding

Auto-scaling Continuous Memory for GUI Agent

FIRE-Bench: Evaluating Research Agents on the Rediscovery of Scientific Insights

Towards General Continuous Memory for Vision-Language Models

Activation Control for Efficiently Eliciting Long Chain-of-thought Ability of Language Models

Exploring the Design Space of Visual Context Representation in Video MLLMs

LESS IS MORE: HIGH-VALUE DATA SELECTION FOR VISUAL INSTRUCTION TUNING

Extracting and Transferring Abilities For Building Multi-lingual Ability-enhanced Large Language Models