LGCRFeb 20, 2025

Fundamental Limitations in Pointwise Defences of LLM Finetuning APIs

arXiv:2502.14828v212 citationsh-index: 9
Originality Highly original
AI Analysis

This work highlights a critical security vulnerability in current LLM fine-tuning defenses, posing risks for developers and users reliant on API safeguards.

The paper demonstrates that pointwise detection defenses for LLM fine-tuning APIs are fundamentally limited, as they constructed attacks using only benign samples that evade detection and successfully elicited harmful answers from the OpenAI API.

LLM developers have imposed technical interventions to prevent fine-tuning misuse attacks, attacks where adversaries evade safeguards by fine-tuning the model using a public API. Previous work has established several successful attacks against specific fine-tuning API defences. In this work, we show that defences of fine-tuning APIs that seek to detect individual harmful training or inference samples ('pointwise' detection) are fundamentally limited in their ability to prevent fine-tuning attacks. We construct 'pointwise-undetectable' attacks that repurpose entropy in benign model outputs (e.g. semantic or syntactic variations) to covertly transmit dangerous knowledge. Our attacks are composed solely of unsuspicious benign samples that can be collected from the model before fine-tuning, meaning training and inference samples are all individually benign and low-perplexity. We test our attacks against the OpenAI fine-tuning API, finding they succeed in eliciting answers to harmful multiple-choice questions, and that they evade an enhanced monitoring system we design that successfully detects other fine-tuning attacks. We encourage the community to develop defences that tackle the fundamental limitations we uncover in pointwise fine-tuning API defences.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes