Bei Yu

SWINGARENA: Adversarial Programming Arena for Long-context GitHub Issue Solving

CircuitNet 3.0: A Multi-Modal Dataset with Task-Oriented Augmentation for AI-Driven Circuit Design

MOSS: Efficient and Accurate FP8 LLM Training with Microscaling and Automatic Scaling

Stratified GRPO: Handling Structural Heterogeneity in Reinforcement Learning of LLM Search Agents

Architect of the Bits World: Masked Autoregressive Modeling for Circuit Generation Guided by Truth Table

VisionReasoner: Unified Reasoning-Integrated Visual Perception via Reinforcement Learning

RePlan: Reasoning-Guided Region Planning for Complex Instruction-Based Image Editing

DreamOmni2: Multimodal Instruction-based Editing and Generation

ToTRL: Unlock LLM Tree-of-Thoughts Reasoning Potential through Puzzles Solving

GAKD: Generative Adversarial Knowledge Distillation For Large Language Models

Grove MoE: Towards Efficient and Superior MoE LLMs with Adjugate Experts

From Noisy Traces to Stable Gradients: Bias--Variance Optimized Preference Optimization for Aligning Large Reasoning Models

Angular and Shell-Aware Deep Potential Energy Model for Molecular Dynamics

MGM-Omni: Scaling Omni LLMs to Personalized Long-Horizon Speech

Generative Distribution Distillation

Don’t Lose the Thread: Empowering Long-Horizon LLM Agents with Cognitive Resource Self-Allocation

One-Token Rollout: Guiding Supervised Fine-Tuning of LLMs with Policy Gradient

FuseGPT: Prune-and-Fuse Knowledge Redistribution for Efficient Transformers

Analytical Restructuring of Feed-Forward Networks for Accelerated LLM Inference

Leveraging Instruction Language Model to Generate Vectorized RISC-V Tensor Programs

Multi-Scale Group Relative Policy Optimization for Large Language Models

DC-LLM: HARDWARE-FRIENDLY LLM WEIGHT COMPRESSION VIA DYNAMIC LINEAR COMBINATION

VisionThink: Smart and Efficient Vision Language Model via Reinforcement Learning

Bridging Layout and RTL: Knowledge Distillation based Timing Prediction

On-Policy Optimization with Group Equivalent Preference for Multi-Programming Language Understanding

PermLLM: Learnable Channel Permutation for N:M Sparse Large Language Models

DLoFT: Gradient-Decoupled Fine-Tuning for Generalizable Long Chain-of-Thought Reasoning

Circuit Representation Learning with Masked Gate Modeling and Verilog-AIG Alignment

TGDPO: Harnessing Token-Level Reward Guidance for Enhancing Direct Preference Optimization

MoreauPruner: Robust Structured Pruning of Large Language Models against Weight Perturbations

Effective LLM Knowledge Learning Requires Rethinking Generalization