Xuehai Pan

Peking University PhD student known for Safe-RLHF, BeaverTails, and core engineering on the OmniSafe and SafeRLHF libraries.

Role: grad-student
Currently at: Peking University
Twitter: twitter.com/XuehaiPan
GitHub: github.com/XuehaiPan
Scholar: scholar.google.com/citations
Papers: 4

Cite

Notes

Only stored in your browser.

Attribution

Affiliations & profile: scholar.google.com/citations

Attribution policy →

4papers·1tool contribs

Authored papers

Kimi k1.5: Scaling Reinforcement Learning with LLMs

arXiv 2025

2025

BeaverTails: Towards Improved Safety Alignment of LLM via a Human-Preference Dataset

NeurIPS

2023

Baichuan 2: Open Large-scale Language Models

arXiv 2023

2023

Safe RLHF: Safe Reinforcement Learning from Human Feedback

arXiv 2023

2023

Tool contributions

PKU-SafeRLHF

PKU-Alignment

Peking University's dual-axis safety + helpfulness preference dataset with explicit harm-category labels, designed for Safe RLHF training.

PreferenceSafetyJailbreak ResistanceInstruction Following

Affiliations

Currently at

Peking University

grad-student · university lab

Frequent co-authors

from 4 papers

Jiaming Ji

researcher

3 shared papers

Mickel Liu

researcher

3 shared papers

Ruiyang Sun

researcher

3 shared papers

Ce Bian

researcher

2 shared papers

Juntao Dai

researcher

2 shared papers

Yaodong Yang

professor

2 shared papers

Yizhou Wang

professor

Aiyuan Yang

Angang Du

Bin Xiao