FOREVER: Forgetting Curve-Inspired Memory Replay for Language Model Continual Learning

TL;DR

FOREVER introduces a memory replay framework for language models that aligns replay timing with the model's internal learning progress, inspired by human forgetting curves, to reduce catastrophic forgetting.

Contribution

It proposes a novel model-centric replay scheduling method based on forgetting curves, improving continual learning performance for large language models.

Findings

FOREVER outperforms existing methods on three benchmarks.

It effectively reduces catastrophic forgetting across models from 0.6B to 13B parameters.

The approach adapts replay intervals based on model evolution rather than fixed steps.

Abstract

Continual learning (CL) for large language models (LLMs) aims to enable sequential knowledge acquisition without catastrophic forgetting. Memory replay methods are widely used for their practicality and effectiveness, but most rely on fixed, step-based heuristics that often misalign with the model's actual learning progress, since identical training steps can result in varying degrees of parameter change. Motivated by recent findings that LLM forgetting mirrors the Ebbinghaus human forgetting curve, we propose FOREVER (FORgEtting curVe-inspired mEmory Replay), a novel CL framework that aligns replay schedules with a model-centric notion of time. FOREVER defines model time using the magnitude of optimizer updates, allowing forgetting curve-inspired replay intervals to align with the model's internal evolution rather than raw training steps. Building on this approach, FOREVER incorporates a forgetting curve-based replay scheduler to determine when to replay and an intensity-aware regularization mechanism to adaptively control how to replay. Extensive experiments on three CL benchmarks and models ranging from 0.6B to 13B parameters demonstrate that FOREVER consistently mitigates catastrophic forgetting.