Revealing the Parametric Knowledge of Language Models: A Unified Framework for Attribution Methods

arXiv:2404.18655v1 Announce Type: new
Abstract: Language Models (LMs) acquire parametric knowledge from their training process, embedding it within their weights. The increasing scalability of LMs, however, poses significant challenges for understanding a model's inner workings and further for updating or correcting this embedded knowledge without the significant cost of retraining. This underscores the importance of unveiling exactly what knowledge is stored and its association with specific model components. Instance Attribution (IA) and Neuron Attribution (NA) offer insights into this training-acquired …

abstract arxiv attribution challenges cs.ai cs.cl embedded embedding framework however knowledge language language models lms parametric process scalability training type understanding