SoFA: Shielded On-the-fly Alignment via Priority Rule Following
This addresses the challenge of aligning LLMs with varied human preferences and standards, representing a novel paradigm rather than an incremental improvement.
The paper tackles the alignment problem in Large Language Models by introducing priority rule following, a paradigm that prioritizes rules over user instructions to adapt to diverse human values and regulatory standards, with experiments showing it effectively minimizes misalignments using only one general rule and adapts smoothly to unseen rules.
The alignment problem in Large Language Models (LLMs) involves adapting them to the broad spectrum of human values. This requirement challenges existing alignment methods due to diversity of preferences and regulatory standards. This paper introduces a novel alignment paradigm, priority rule following, which defines rules as the primary control mechanism in each dialog, prioritizing them over user instructions. Our preliminary analysis reveals that even the advanced LLMs, such as GPT-4, exhibit shortcomings in understanding and prioritizing the rules. Therefore, we present PriorityDistill, a semi-automated approach for distilling priority following signals from LLM simulations to ensure robust rule integration and adherence. Our experiments show that this method not only effectively minimizes misalignments utilizing only one general rule but also adapts smoothly to various unseen rules, ensuring they are shielded from hijacking and that the model responds appropriately.