AI & ML interests
A research group from Zhejiang University working on KG and NLP Technologies
Recent Activity
Papers
TokenPilot: Cache-Efficient Context Management for LLM Agents
Unsupervised Skill Discovery for Agentic Data Analysis
Benchmarking Agentic Workflow Generation
Data and models for Mol-Instructions.
A Multi-Modal AI Copilot for Single-Cell Analysis with Instruction Following
-
zjunlp/Qwen2.5-9B-Instruct-CBM-RD
9B β’ Updated β’ 2 β’ 1 -
zjunlp/Qwen2.5-9B-Instruct-CBM-CD
9B β’ Updated β’ 8 β’ 2 -
When Should Models Change Their Minds? Contextual Belief Management in Large Language Models
Paper β’ 2605.30219 β’ Published β’ 26 -
zjunlp/BeliefTrackDataset
Updated β’ 1.51k β’ 4
A Large-Scale Multimodal Ocean Corpus for Foundation Models
-
OceanPile: A Large-Scale Multimodal Ocean Corpus for Foundation Models
Paper β’ 2605.00877 β’ Published β’ 15 -
zjunlp/OceanCorpus
Viewer β’ Updated β’ 38.4k β’ 2.27k β’ 5 -
zjunlp/OceanBenchmark
Viewer β’ Updated β’ 1.02k β’ 234 β’ 1 -
zjunlp/OceanInstruction
Viewer β’ Updated β’ 400 β’ 2.32k β’ 3
-
zjunlp/OceanGPT-basic-30B-A3B-Instruct
Text Generation β’ 31B β’ Updated β’ 9 β’ 1 -
zjunlp/OceanGPT-basic-30B-A3B-Thinking
Text Generation β’ 31B β’ Updated β’ 6 β’ 1 -
zjunlp/OceanGPT-basic-4B-Instruct
Text Generation β’ 4B β’ Updated β’ 29 β’ β’ 1 -
zjunlp/OceanGPT-basic-4B-Thinking
Text Generation β’ 4B β’ Updated β’ 21 β’ β’ 1
-
zjunlp/DataMind-Analysis-Qwen2.5-7B
Text Generation β’ 8B β’ Updated β’ 10 β’ 2 -
zjunlp/DataMind-Analysis-Qwen2.5-14B
Text Generation β’ 15B β’ Updated β’ 5 β’ 2 -
Why Do Open-Source LLMs Struggle with Data Analysis? A Systematic Empirical Study
Paper β’ 2506.19794 β’ Published β’ 8 -
zjunlp/DataMind-Analysis-SFT-Data
Viewer β’ Updated β’ 2.82k β’ 52 β’ 1
model adapter weight and train data for "Agent Planning with World Knowledge Model"
Merged models generated for the model kinship paper.
-
zjunlp/Qwen2.5-9B-Instruct-CBM-RD
9B β’ Updated β’ 2 β’ 1 -
zjunlp/Qwen2.5-9B-Instruct-CBM-CD
9B β’ Updated β’ 8 β’ 2 -
When Should Models Change Their Minds? Contextual Belief Management in Large Language Models
Paper β’ 2605.30219 β’ Published β’ 26 -
zjunlp/BeliefTrackDataset
Updated β’ 1.51k β’ 4
A Large-Scale Multimodal Ocean Corpus for Foundation Models
-
OceanPile: A Large-Scale Multimodal Ocean Corpus for Foundation Models
Paper β’ 2605.00877 β’ Published β’ 15 -
zjunlp/OceanCorpus
Viewer β’ Updated β’ 38.4k β’ 2.27k β’ 5 -
zjunlp/OceanBenchmark
Viewer β’ Updated β’ 1.02k β’ 234 β’ 1 -
zjunlp/OceanInstruction
Viewer β’ Updated β’ 400 β’ 2.32k β’ 3
-
zjunlp/OceanGPT-basic-30B-A3B-Instruct
Text Generation β’ 31B β’ Updated β’ 9 β’ 1 -
zjunlp/OceanGPT-basic-30B-A3B-Thinking
Text Generation β’ 31B β’ Updated β’ 6 β’ 1 -
zjunlp/OceanGPT-basic-4B-Instruct
Text Generation β’ 4B β’ Updated β’ 29 β’ β’ 1 -
zjunlp/OceanGPT-basic-4B-Thinking
Text Generation β’ 4B β’ Updated β’ 21 β’ β’ 1
-
zjunlp/DataMind-Analysis-Qwen2.5-7B
Text Generation β’ 8B β’ Updated β’ 10 β’ 2 -
zjunlp/DataMind-Analysis-Qwen2.5-14B
Text Generation β’ 15B β’ Updated β’ 5 β’ 2 -
Why Do Open-Source LLMs Struggle with Data Analysis? A Systematic Empirical Study
Paper β’ 2506.19794 β’ Published β’ 8 -
zjunlp/DataMind-Analysis-SFT-Data
Viewer β’ Updated β’ 2.82k β’ 52 β’ 1
Benchmarking Agentic Workflow Generation
model adapter weight and train data for "Agent Planning with World Knowledge Model"
Merged models generated for the model kinship paper.
Data and models for Mol-Instructions.
A Multi-Modal AI Copilot for Single-Cell Analysis with Instruction Following