NeuroActiSep: Detecting Factual Hallucinations from Feed-Forward Neurons in a Single Pass
This work proposes a method to rank feed-forward neurons at the final prompt token using a custom neuron selection dataset, and transfers the selected neuron identities to train hallucination classifiers on other factual question answering datasets.
Ali Derogar Odolou, Reza Nazari, Mostafa Salehi
· 0 citations