Xin Jiang

Expert Merging: Model Merging with Unsupervised Expert Alignment and Importance-Guided Layer Chunking

ToolACE-MT: Non-Autoregressive Generation for Agentic Multi-Turn Interaction

FlatQuant: Flatness Matters for LLM Quantization

ToolACE: Winning the Points of LLM Function Calling

Beyond Autoregression: Discrete Diffusion for Complex Reasoning and Planning

Bridging and Modeling Correlations in Pairwise Data for Direct Preference Optimization

Forewarned is Forearmed: Harnessing LLMs for Data Synthesis via Failure-induced Exploration

RevisEval: Improving LLM-as-a-Judge via Response-Adapted References

DAPE V2: Process Attention Score as Feature Map for Length Extrapolation

Subtle Errors Matter: Preference Learning via Error-injected Self-editing

SepLLM: Accelerate Large Language Models by Compressing One Segment into One Separator

FlatQuant: Flatness Matters for LLM Quantization

Implicit Search via Discrete Diffusion: A Study on Chess

Mixture of insighTful Experts (MoTE): The Synergy of Thought Chains and Expert Mixtures in Safety Self-Alignment

Efficient Multi-modal Large Language Models via Visual Token Grouping