Xiaoyu Wen

XiaoyuWen's activity

• All
• Models
• Datasets
• Spaces
• Buckets
• Papers
• Collections
• Community
• Posts
• Upvotes
• Likes
• Articles

updated a collection about 2 months ago

MAGIC

Collection

ICML2026 paper: "MAGIC: A Co-Evolving Attacker–Defender Adversarial Game for Robust LLM Safety" • 7 items • Updated Aug 26 • 3

updated 3 models about 2 months ago

XiaoyuWen/MAGIC-attacker-sft-Qwen2.5-14B-Instruct

15B • Updated Aug 26 • 6

XiaoyuWen/MAGIC-attacker-sft-Qwen2.5-7B-Instruct

8B • Updated Aug 26 • 17

XiaoyuWen/MAGIC-attacker-sft-Llama3.1-8B-Instruct

8B • Updated Aug 26 • 9

published 3 models about 2 months ago

XiaoyuWen/MAGIC-attacker-sft-Qwen2.5-7B-Instruct

8B • Updated Aug 26 • 17

XiaoyuWen/MAGIC-attacker-sft-Qwen2.5-14B-Instruct

15B • Updated Aug 26 • 6

XiaoyuWen/MAGIC-attacker-sft-Llama3.1-8B-Instruct

8B • Updated Aug 26 • 9

updated a collection about 2 months ago

TRACE

Collection

Paper: "Not All Turns Matter: Credit Assignment for Multi-Turn Jailbreaking." • 2 items • Updated Aug 25

updated a model about 2 months ago

XiaoyuWen/TRACE

Reinforcement Learning • Updated Aug 25

published a model about 2 months ago

XiaoyuWen/TRACE

Reinforcement Learning • Updated Aug 25

updated a collection about 2 months ago

TRACE

Collection

Paper: "Not All Turns Matter: Credit Assignment for Multi-Turn Jailbreaking." • 2 items • Updated Aug 25

updated a model about 2 months ago

XiaoyuWen/TRACE-Mix-Qwen2.5-3B-Instruct

Text Generation • 3B • Updated Aug 25 • 661 • 2

published a model about 2 months ago

XiaoyuWen/TRACE-Mix-Qwen2.5-3B-Instruct

Text Generation • 3B • Updated Aug 25 • 661 • 2

updated a dataset 5 months ago

XiaoyuWen/PIA-Persona-Dataset

Viewer • Updated May 10 • 35 • 15

updated a collection 5 months ago

PIA

Collection

ICML2026 paper: "Disentangling Intent from Role: Adversarial Self-Play for Persona-Invariant Safety Alignment" • 2 items • Updated May 10 • 1

upvoted a collection 5 months ago

PIA

Collection

ICML2026 paper: "Disentangling Intent from Role: Adversarial Self-Play for Persona-Invariant Safety Alignment" • 2 items • Updated May 10 • 1

updated a model 5 months ago

XiaoyuWen/PIA

Updated May 10