Saved in:
Bibliographic Details
Main Authors: Kapusuzoglu, Berkcan, Chakraborty, Supriyo, Ni, Renkun, Rawls, Stephen, Sahu, Sambit
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2511.08500
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912702267392000
author Kapusuzoglu, Berkcan
Chakraborty, Supriyo
Ni, Renkun
Rawls, Stephen
Sahu, Sambit
author_facet Kapusuzoglu, Berkcan
Chakraborty, Supriyo
Ni, Renkun
Rawls, Stephen
Sahu, Sambit
contents Large language models (LLMs) adapted to financial domains often suffer from catastrophic forgetting of general reasoning capabilities essential for customer interactions and complex financial analysis. We introduce Selective Parameter Evaluation and Restoration via Model Merging (SPEAR-MM), a practical framework that preserves critical capabilities while enabling domain adaptation. Our method approximates layer-wise impact on external benchmarks through post-hoc analysis, then selectively freezes or restores transformer layers via spherical interpolation merging. Applied to LLaMA-3.1-8B for financial tasks, SPEAR-MM achieves 91.2% retention of general capabilities versus 69.7% for standard continual pretraining, while maintaining 94% of domain adaptation gains. The approach provides interpretable trade-off control and reduces computational costs by 90% crucial for resource-constrained financial institutions.
format Preprint
id arxiv_https___arxiv_org_abs_2511_08500
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle SPEAR-MM: Selective Parameter Evaluation and Restoration via Model Merging for Efficient Financial LLM Adaptation
Kapusuzoglu, Berkcan
Chakraborty, Supriyo
Ni, Renkun
Rawls, Stephen
Sahu, Sambit
Computation and Language
Artificial Intelligence
Machine Learning
Spectral Theory
Large language models (LLMs) adapted to financial domains often suffer from catastrophic forgetting of general reasoning capabilities essential for customer interactions and complex financial analysis. We introduce Selective Parameter Evaluation and Restoration via Model Merging (SPEAR-MM), a practical framework that preserves critical capabilities while enabling domain adaptation. Our method approximates layer-wise impact on external benchmarks through post-hoc analysis, then selectively freezes or restores transformer layers via spherical interpolation merging. Applied to LLaMA-3.1-8B for financial tasks, SPEAR-MM achieves 91.2% retention of general capabilities versus 69.7% for standard continual pretraining, while maintaining 94% of domain adaptation gains. The approach provides interpretable trade-off control and reduces computational costs by 90% crucial for resource-constrained financial institutions.
title SPEAR-MM: Selective Parameter Evaluation and Restoration via Model Merging for Efficient Financial LLM Adaptation
topic Computation and Language
Artificial Intelligence
Machine Learning
Spectral Theory
url https://arxiv.org/abs/2511.08500