Emergent and Predictable Memorization in Large Language Models

Understand

Memorization, or the tendency of large language models (LLMs) to output entire sequences from their training data verbatim, is a key concern for safely deploying language models.

• In particular, it is vital to minimize a model's memorization of sensitive datapoints such as those containing personal identifiable information (PII).
• The prevalence of such undesirable memorization can pose issues for model trainers, and may even require discarding an otherwise functional model.
• We therefore seek to predict which sequences will be memorized before a large model's full train-time by extrapolating the memorization behavior of lower-compute trial runs.