Weight space Detection of Backdoors in LoRA Adapters

arXiv:2602.15195v2 Announce Type: replace Abstract: LoRA adapters let users fine-tune large language models (LLMs) efficiently. However, LoRA adapters are shared through open repositories like Hugging Face Hub \citep{huggingface_hub_docs}, making them vulnerable to backdoor at...

🔗 Read more: https://arxiv.org/abs/2602.15195

#News #Psychology #Math #Policy #AI #Space #Academic