Scholar

Pengkai Wang

Google Scholar ID: v8QXxpIAAAAJ

Zhejiang University | HK Polytechnic University

control & optimizationAI for ScienceLLM-post training

Google Scholar↗

Citations & Impact

All-time

Citations

31

H-index

3

i10-index

1

Publications

7

Co-authors

0

Contact

No contact links provided.

Publications

8 items

Not All Disagreement Is Learnable: Token Teachability in On-Policy Distillation

2026

Cited

0

E-PMQ: Expert-Guided Post-Merge Quantization with Merged-Weight Anchoring

2026

Cited

0

Discovering Physical Directions in Weight Space: Composing Neural PDE Experts

2026

Cited

0

FeatCal: Feature Calibration for Post-Merging Models

2026

Cited

0

Geometry Conflict: Explaining and Controlling Forgetting in LLM Continual Post-Training

2026

Cited

0

An Efficient Graph-Transformer Operator for Learning Physical Dynamics with Manifolds Embedding

2025

Cited

0

InfiMed-ORBIT: Aligning LLMs on Open-Ended Complex Tasks via Rubric-Based Incremental Training

2025

Cited

0

Learning Self-Growth Maps for Fast and Accurate Imbalanced Streaming Data Clustering

2024

Cited

1

Resume (English only)

Co-authors

0 total

Co-authors: 0 (list not available)