Travis Lelle's picture

Open to Collab

Travis Lelle PRO

Travis-ML

·

https://www.travisML.ai

AI & ML interests

Adversarial ML and AI - LoRA Adapter Backdoors - AI-Security in general

Recent Activity

published an article about 16 hours ago

How I Backed Into AI Research by Trying to Poison My Own Model

authored a paper 1 day ago

Token-Level Generalization in LoRA Adapter Backdoors: Attack Characterization and Behavioral Detection

commentedon a paper 3 days ago

Token-Level Generalization in LoRA Adapter Backdoors: Attack Characterization and Behavioral Detection

View all activity

Organizations

published an article about 16 hours ago

Article

How I Backed Into AI Research by Trying to Poison My Own Model

Travis-ML

•

about 16 hours ago

authored a paper 1 day ago

Token-Level Generalization in LoRA Adapter Backdoors: Attack Characterization and Behavioral Detection

Paper • 2605.30189 • Published 6 days ago • 5

commented a paper 3 days ago

Token-Level Generalization in LoRA Adapter Backdoors: Attack Characterization and Behavioral Detection

Paper • 2605.30189 • Published 6 days ago • 5 •

updated a collection 3 days ago

LoRA Adapter Backdoor Research (arXiv:2605.30189)

Companion artifacts for "LoRA Adapter Backdoor Research" (arXiv:2605.30189): intentionally backdoored LoRA adapters and the poisoned training datasets • 34 items • Updated 3 days ago