Ningyu Zhang

Scaling Generalist Data-Analytic Agents

LightMem: Lightweight and Efficient Memory-Augmented Generation

Scaling Agents via Continual Pre-training

InnoGym: Benchmarking the Innovation Potential of AI Agents

Towards Personalized Deep Research: Benchmarks and Evaluations

Towards General Agentic Intelligence via Environment Scaling

OceanGym: A Benchmark Environment for Underwater Embodied Agents

Pushing LLMs to Their Logical Reasoning Bound: The Role of Data Reasoning Intensity

Safetylock: Guarding LLM againt FuneTuning Risks with Efficient Inference-time Addon

UniErase: Towards Balanced and Precise Unlearning in Language Models

AnyEdit: Edit Any Knowledge Encoded in Language Models

Two Experts Are All You Need for Steering Thinking: Reinforcing Cognitive Effort in MoE Reasoning Models Without Additional Training

Benchmarking Agentic Workflow Generation

MLLM can see? Dynamic Correction Decoding for Hallucination Mitigation

Exploring Model Kinship for Merging Large Language Models

Locking Down the Finetuned LLMs Safety

Symbolic Learning Enables Self-Evolving Agents