LoRA as Oracle
Practitioners increasingly deploy neural networks they did not train, and must audit them after the fact for hidden backdoors, without the training pipeline, the poisoned data, or knowledge of any trigger. We introduce a low-rank auditing lens built on a single observation: what a model has internalized and how it behaves are distinct axes that can diverge. Fitting a small low-rank adapter toward a hypothesis and reading the geometry of the resulting update, its energy relative to, and its alignment with, the frozen weights, measures internalization directly, independently of the model's output behavior. The lens earns its value where the two axes diverge: a backdoor is malicious internalization, a learned trigger-to-target shortcut, that behavioral auditing can miss. Reading it, we identify the backdoor's target class label-free and without any triggered data when the backdoor leaves a legible internalization signature, then erase the identified shortcut within the same low-rank subspace, uniquely coupling detection confidence to the size of the repair. Across four datasets and four architectures, our lens is the most consistent target identifier among defenses in its threat model; its rank-r repair removes backdoors while preserving clean accuracy where full-model baselines collapse it, at orders-of-magnitude lower parameter and memory cost; and it is the only method that both erases the backdoor and leaves a benign model intact, because it can decline to act when uncertain.
Code (0)
등록된 구현이 없습니다.
Similar Papers 제목 키워드 기반
Policy Gradient as a Proxy for Dynamic Oracles in Constituency Parsing
Dynamic oracles provide strong supervision for training constituency parsers with exploration, but must be custom defined for a given parser's transition system. We explore using a policy gradient method as a parser-agno…
Constituency ParsingGeneric Oracles for Structured Prediction
When learned without exploration, local models for structured prediction tasks are subject to exposure bias and cannot be trained without detailed guidance. Active Imitation Learning (AIL), also known in NLP as Dynamic O…
Imitation LearningPredictionStructured PredictionSome Supervision Required: Incorporating Oracle Policies in Reinforcement Learning via Epistemic Uncertainty Metrics
An inherent problem of reinforcement learning is performing exploration of an environment through random actions, of which a large portion can be unproductive. Instead, exploration can be improved by initializing the lea…
reinforcement-learningReinforcement LearningReinforcement Learning (RL)Active Policy Improvement from Multiple Black-box Oracles
Reinforcement learning (RL) has made significant strides in various complex domains. However, identifying an effective policy via RL often necessitates extensive exploration. Imitation learning aims to mitigate this issu…
Imitation LearningReinforcement Learning (RL)Contextual Bandits in Payment Processing: Non-uniform Exploration and Supervised Learning at Adyen
Uniform random exploration in decision-making systems supports off-policy learning via supervision but incurs high regret, making it impractical for many applications. Conversely, non-uniform exploration offers better im…
Multi-Armed Banditsregression