Safe fine-tuning practices, alignment preservation, and data curation
Newest first · 4 reviewed on this page
Search insteadWenjuan Li, Yitao Liu, Runze Chen +1
Background: Fine-tuning is central to adapting pre-trained Large Language Models (LLMs) to downstream tasks, but its reliance on training data, parameter updates, and reusable components opens entry points for attackers. Threats have evolved from data poisoning and weight…
Srikanth Malla, Chiho Choi, Joon Choi
Post-hoc safety training (RLHF, DPO) is the dominant way to align large language models, yet jailbreaks (Zou et al., 2023b), fine-tuning attacks (Qi et al., 2024), and activation-space probes (Arditi et al., 2024) keep recovering the behaviors it was meant to remove. We give…
Herbert Woisetschläger, Alexander Isenko, Shiqiang Wang +2
Examines federated learning approaches for fine-tuning LLMs on edge devices, analyzing privacy guarantees, communication efficiency, and security trade-offs.
Alexander Wan, Eric Wallace, Sheng Shen +1
Shows that adversaries can insert poisoned examples into instruction-tuning datasets, causing models to generate targeted outputs for attacker-chosen triggers.
Simon Lermen, Charlie Rogers-Smith, Jeffrey Ladish
Shows that LoRA fine-tuning with as few as 100 examples can remove safety guardrails from Llama 2-Chat, raising concerns about fine-tuning access to aligned models.
Xianjun Yang, Xiao Wang, Qi Zhang +4
Shows that RLHF can introduce shadow alignment where models exhibit harmful behaviors not present in the base model.