Forgetting in Language Models: Capacity, Optimization, and Self-Generated Replay
Models trained on a new task typically degrade on prior tasks, a phenomenon known as forgetting. Traditionally, mitigating forgetting has required replaying stored exemplars from prior tasks, which is often impractical. By contrast, language models can sample from their own training distribution, and we show that these self-generated samples serve as effective replay data, nearly eliminating forgetting. We find that forgetting nonetheless persists when the model has little remaining capacity: models pretrained close to saturation cannot absorb new information without overwriting prior knowledge. When capacity is not the limiting factor, low learning rates reduce forgetting but require substantially more training steps. Replay breaks this tradeoff, enabling fast, high-learning-rate finetuning without forgetting.
Code (0)
등록된 구현이 없습니다.
Similar Papers 제목 키워드 기반
Talking to Yourself: Defying Forgetting in Large Language Models
Catastrophic forgetting remains a major challenge when fine-tuning large language models (LLMs) on narrow, task-specific data, often degrading their general knowledge and reasoning abilities. We propose SA-SFT, a lightwe…
General KnowledgeUNDIAL: Self-Distillation with Adjusted Logits for Robust Unlearning in Large Language Models
Mitigating the retention of sensitive or private information in large language models is essential for enhancing privacy and safety. Existing unlearning methods, like Gradient Ascent and Negative Preference Optimization,…
Natural Language Understandingparameter-efficient fine-tuningSCM: Sleep-Consolidated Memory with Algorithmic Forgetting for Large Language Models
We present SCM (Sleep-Consolidated Memory), a research preview of a memory architecture for large language models that draws on neuroscientific principles to address a fundamental limitation in current systems: the absen…
Graceful Forgetting in Generative Language Models
Recently, the pretrain-finetune paradigm has become a cornerstone in various deep learning areas. While in general the pre-trained model would promote both effectiveness and efficiency of downstream tasks fine-tuning, st…
Associative Memory for Non-Stationary Environments: A Self-Sizing Generalization of Hopfield Networks
The Hopfield network made associative memory (AM) the model system of neural computation, but it solves the problem only for a \emph{stationary} world: a fixed set of memories, stored once into frozen weights. Real envir…