42 70 103

Ningyu Zhang

Ningyu

https://person.zju.edu.cn/en/ningyu

zxlzr

AI & ML interests

NLP, Knowledge Editing

Recent Activity

upvoted a paper 9 days ago

Spatial Knowledge Graph-Guided Multimodal Synthesis

commented on a paper 9 days ago

Spatial Knowledge Graph-Guided Multimodal Synthesis

authored a paper 10 days ago

Beyond Prompt Engineering: Robust Behavior Control in LLMs via Steering Target Atoms

View all activity

Organizations

Ningyu's activity

upvoted a paper 9 days ago

Spatial Knowledge Graph-Guided Multimodal Synthesis

Paper • 2505.22633 • Published 9 days ago • 3

commented a paper 9 days ago

Spatial Knowledge Graph-Guided Multimodal Synthesis

Paper • 2505.22633 • Published 9 days ago • 3 •

authored a paper 10 days ago

Beyond Prompt Engineering: Robust Behavior Control in LLMs via Steering Target Atoms

Paper • 2505.20322 • Published 14 days ago • 14

upvoted a paper 10 days ago

Beyond Prompt Engineering: Robust Behavior Control in LLMs via Steering Target Atoms

Paper • 2505.20322 • Published 14 days ago • 14

commented a paper 10 days ago

Beyond Prompt Engineering: Robust Behavior Control in LLMs via Steering Target Atoms

Paper • 2505.20322 • Published 14 days ago • 14 •

liked a model 15 days ago

zjunlp/OceanGPT-basic-7B-v0.3

Text Generation • Updated 25 days ago • 30 • 2

commented on Mixture of Tunable Experts - Behavior Modification of DeepSeek-R1 at Inference Time 16 days ago

This work is really impressive! I just discovered this blog. While researching steering methods, we also found that it's possible to change the model's language behavior. We've integrated some existing steering methods into EasyEdit2:
https://github.com/zjunlp/EasyEdit/blob/main/README_2.md

Here's a notebook related to language control:
https://github.com/zjunlp/EasyEdit/blob/main/tutorial-notebooks/EasyEdit2_Example_CAA_translate.ipynb

upvoted an article 16 days ago

Article

Mixture of Tunable Experts - Behavior Modification of DeepSeek-R1 at Inference Time

and 4 others •

Feb 18

• 33

authored 7 papers 17 days ago

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

Paper • 2505.14681 • Published 17 days ago • 9

upvoted a paper 17 days ago

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

Paper • 2505.14681 • Published 17 days ago • 9

commented a paper 17 days ago

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

Paper • 2505.14681 • Published 17 days ago • 9 •

liked a model 17 days ago

zjunlp/OceanGPT-basic-8B

Text Generation • Updated 18 days ago • 21 • 1

updated a model 23 days ago

zjunlp/OceanGPT-o-7B

Text Generation • Updated 23 days ago • 36 • 1

authored a paper 24 days ago

Knowledge Augmented Complex Problem Solving with Large Language Models: A Survey

Paper • 2505.03418 • Published May 6 • 8