5 papers
Verification of the Implicit World Model in a Generative Model via Adversarial Sequences
András Balogh, Márk Jelasity
Generative sequence models are typically trained on sample sequences from natural or formal languages. It is a crucial question whether -- or to what extent -- sample-based trainin…
Detecting Semantic Backdoors in a Mystery Shopping Scenario
Arpad Berta, Gabor Danner, Istvan Hegedus +1
Detecting semantic backdoors in classification models--where some classes can be activated by certain natural, but out-of-distribution inputs--is an important problem that has rece…
On the Brittleness of LLMs: A Journey around Set Membership
Lea Hergert, Gábor Berend, Mario Szegedy +2
Large language models (LLMs) achieve superhuman performance on complex reasoning tasks, yet often fail on much simpler problems, raising concerns about their reliability and interp…
No Soundness in the Real World: On the Challenges of the Verification of Deployed Neural Networks
Attila Szász, Balázs Bánhelyi, Márk Jelasity
The ultimate goal of verification is to guarantee the safety of deployed neural networks. Here, we claim that all the state-of-the-art verifiers we are aware of fail to reach this…
How not to Stitch Representations to Measure Similarity: Task Loss Matching versus Direct Matching
András Balogh, Márk Jelasity
Measuring the similarity of the internal representations of deep neural networks is an important and challenging problem. Model stitching has been proposed as a possible approach,…