Learn Before
Mitigation Strategies for Catastrophic Forgetting
To counteract catastrophic forgetting during fine-tuning, several methods can be used. A common practical technique is to include data from the original tasks in the fine-tuning process, promoting a more diverse training set. More specialized approaches developed for continual learning, such as experience replay and elastic weight consolidation, can also be applied.
0
1
Tags
Ch.2 Generative Models - Foundations of Large Language Models
Foundations of Large Language Models
Computing Sciences
Ch.1 Pre-training - Foundations of Large Language Models
Foundations of Large Language Models Course
Related
Mitigation Strategies for Catastrophic Forgetting
A team starts with a large language model that is highly proficient at a wide range of general language tasks, including text summarization, translation, and question-answering. They then fine-tune this model exclusively on a new, highly specialized dataset of legal document summaries. After this training, the model becomes excellent at summarizing legal documents but is now significantly worse at performing general translation than it was before. Which phenomenon does this scenario most directly demonstrate?
Diagnosing Performance Degradation in a Fine-Tuned Model
Illustrating a Key Fine-Tuning Challenge