pith. sign in

arxiv: 2407.08699 · v2 · pith:KOTMXGBXnew · submitted 2024-07-11 · 💻 cs.LG

Mitigating Catastrophic Forgetting in Language Transfer via Model Merging

classification 💻 cs.LG
keywords forgettingmodeldomainlanguagemodelsacrossadaptationcatastrophic
0
0 comments X
read the original abstract

As open-weight large language models (LLMs) achieve ever more impressive performances across a wide range of tasks in English, practitioners aim to adapt these models to different languages. However, such language adaptation is often accompanied by catastrophic forgetting of the base model's capabilities, severely limiting the usefulness of the resulting model. We address this issue by proposing Branch-and-Merge (BaM), a new adaptation method based on iteratively merging multiple models, fine-tuned on a subset of the available training data. BaM is based on the insight that this yields lower magnitude but higher quality weight changes, reducing forgetting of the source domain while maintaining learning on the target domain. We demonstrate in an extensive empirical study on Bulgarian and German that BaM can significantly reduce forgetting while matching or even improving target domain performance compared to both standard continued pretraining and instruction finetuning across different model architectures.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. ORBIT: Preserving Foundational Language Capabilities in GenRetrieval via Origin-Regulated Merging

    cs.CL 2026-05 unverdicted novelty 5.0

    ORBIT preserves foundational language capabilities during generative retrieval fine-tuning by using origin-regulated weight averaging to constrain parameter drift beyond a distance threshold.

  2. Model Merging in LLMs, MLLMs, and Beyond: Methods, Theories, Applications and Opportunities

    cs.LG 2024-08 accept novelty 4.0

    The paper introduces a new taxonomy for model merging methods and reviews their applications in LLMs, MLLMs, continual learning, multi-task learning, and other subfields while outlining open challenges.