Understanding Catastrophic Forgetting in Language Models via Implicit Inference

Open in new window