Xuanjing Huang

AgentGym-RL: An Open-Source Framework to Train LLM Agents for Long-Horizon Decision Making via Multi-Turn RL

RECAST: Expanding the Boundaries of LLMs' Complex Instruction Following with Multi-Constraint Data

R-Horizon: How Far Can Your Large Reasoning Model Really Go in Breadth and Depth?

Biologically Plausible Learning via Bidirectional Spike-Based Distillation

Game-RL: Synthesizing Multimodal Verifiable Game Data to Boost VLMs' General Reasoning

EgoNight: Towards Egocentric Vision Understanding at Night with a Challenging Benchmark

Critique-RL: Training Language Models For Critiquing Through Two-Stage Reinforcement Learning

BAPO: Stabilizing Off-Policy Reinforcement Learning for LLMs via Balanced Policy Optimization with Adaptive Clipping

Why Reinforcement Fine-Tuning Enables MLLMs Preserve Prior Knowledge Better: A Data Perspective

Adaptive Curriculum Strategies: Stabilizing Reinforcement Learning for Large Language Models

Unlocking the Essence of Beauty: Advanced Aesthetic Reasoning with Relative-Absolute Policy Optimization

Two Minds Better Than One: Collaborative Reward Modeling for LLM Alignment

WorldPM: Understanding Scaling Patterns in Human Preference Modeling

Search or Think? Rethinking Iterative RAG from An Entropy Perspective

Mitigating Position Bias in Transformers via Layer-Specific Positional Embedding Scaling

MDAR: A Multi-scene Dynamic Audio Reasoning Benchmark

From Scores to Preferences: Redefining MOS Benchmarking for Speech Quality Reward Modeling

Context-Aware Alignment: Adapting Large Language Models to Individual Historical Data

Model Utility Law: Evaluating LLMs beyond Performance via Mechanistically Interpretable Metric

EvaLearn: Quantifying the Learning Capability and Efficiency of LLMs via Sequential Problem Solving

Understanding Parametric and Contextual Knowledge Reconciliation within Large Language Models

Toward Relative Positional Encoding in Spiking Transformers

Pre-Trained Policy Discriminators are General Reward Models

Have the VLMs Lost Confidence? A Study of Sycophancy in VLMs

Domain-RAG: Retrieval-Guided Compositional Image Generation for Cross-Domain Few-Shot Object Detection

Effective Length Extrapolation via Dimension-Wise Positional Embeddings Manipulation

Distill Visual Chart Reasoning Ability from LLMs to MLLMs

RMB: Comprehensively benchmarking reward models in LLM alignment

How Jailbreak Defenses Work and Ensemble? A Mechanistic Investigation

AgentGym: Evaluating and Evolving Large Language Model-based Agents across Diverse Envronments

SPA-VL: A Comprehensive Safety Preference Alignment Dataset for Vision Language Model

Tell Me What You Don't Know: Enhancing Refusal Capabilities of Role-Playing Agents via Representation Space Analysis and Editing

Dendritic Localized Learning: Toward Biologically Plausible Algorithm

Progressive LLM Alignments Using Two-Player Games