Showing cs.CLShow all
3 papers · 1 filter
cs.CL2025
Covering Cracks in Content Moderation: Delexicalized Distant Supervision for Illicit Drug Jargon Detection
Minkyoo Song, Eugene Jang, Jaehan Kim +1
In light of rising drug-related concerns and the increasing role of social media, sales and discussions of illicit drugs have become commonplace online. Social media platforms host…
cs.CL2024
Obliviate: Neutralizing Task-agnostic Backdoors within the Parameter-efficient Fine-tuning Paradigm
Jaehan Kim, Minkyoo Song, Seung Ho Na +1
Parameter-efficient fine-tuning (PEFT) has become a key training strategy for large language models. However, its reliance on fewer trainable parameters poses security risks, such…
cs.CL2024
Claim-Guided Textual Backdoor Attack for Practical Applications
Minkyoo Song, Hanna Kim, Jaehan Kim +2
Recent advances in natural language processing and the increased use of large language models have exposed new security vulnerabilities, such as backdoor attacks. Previous backdoor…