Tübingen Team Proposes Replay on Demand for Continual Pretraining

Researchers from Tübingen and Basel introduced "Replay on Demand" (RoD), a continual pretraining method that dynamically allocates replay data online based on forgetting, balancing adaptation to new domains with retention of old capabilities.

2026-10-02 ~ 2026-10-02 · 2 related posts

• Replay on Demand paper allocates replay data online from model forgetting dynamics — schwarzjn_ · 2026-10-02

1 near-duplicate retellings: schwarzjn_