Meta-Learning at Scale for Large Language Models via Low-Rank Amortized Bayesian Meta-Learning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zhang, Liyi, Snell, Jake, Griffiths, Thomas L.
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915906498592768
author Zhang, Liyi
Snell, Jake
Griffiths, Thomas L.
author_facet Zhang, Liyi
Snell, Jake
Griffiths, Thomas L.
contents Fine-tuning large language models (LLMs) with low-rank adaptation (LoRA) is a cost-effective way to incorporate information from a specific dataset. However, when a problem requires incorporating information from multiple datasets - as in few shot learning - generalization across datasets can be limited, driving up training costs. As a consequence, other approaches such as in-context learning are typically used in this setting. To address this challenge, we introduce an efficient method for adapting the weights of LLMs to multiple distributions, Amortized Bayesian Meta-Learning for LoRA (ABMLL). This method builds on amortized Bayesian meta-learning for smaller models, adapting this approach to LLMs by reframing where local and global variables are defined in LoRA and using a new hyperparameter to balance reconstruction accuracy and the fidelity of task-specific parameters to the global ones. ABMLL supports effective generalization across datasets and scales to large models such as Llama3-8B and Qwen2-7B, outperforming existing methods on the CrossFit and Unified-QA datasets in terms of both accuracy and expected calibration error. We show that meta-learning can also be combined with in-context learning, resulting in further improvements in both these datasets and legal and chemistry applications.
format Preprint
id arxiv_https___arxiv_org_abs_2508_14285
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Meta-Learning at Scale for Large Language Models via Low-Rank Amortized Bayesian Meta-Learning
Zhang, Liyi
Snell, Jake
Griffiths, Thomas L.
Machine Learning
Artificial Intelligence
I.2.7
Fine-tuning large language models (LLMs) with low-rank adaptation (LoRA) is a cost-effective way to incorporate information from a specific dataset. However, when a problem requires incorporating information from multiple datasets - as in few shot learning - generalization across datasets can be limited, driving up training costs. As a consequence, other approaches such as in-context learning are typically used in this setting. To address this challenge, we introduce an efficient method for adapting the weights of LLMs to multiple distributions, Amortized Bayesian Meta-Learning for LoRA (ABMLL). This method builds on amortized Bayesian meta-learning for smaller models, adapting this approach to LLMs by reframing where local and global variables are defined in LoRA and using a new hyperparameter to balance reconstruction accuracy and the fidelity of task-specific parameters to the global ones. ABMLL supports effective generalization across datasets and scales to large models such as Llama3-8B and Qwen2-7B, outperforming existing methods on the CrossFit and Unified-QA datasets in terms of both accuracy and expected calibration error. We show that meta-learning can also be combined with in-context learning, resulting in further improvements in both these datasets and legal and chemistry applications.
title Meta-Learning at Scale for Large Language Models via Low-Rank Amortized Bayesian Meta-Learning
topic Machine Learning
Artificial Intelligence
I.2.7
url https://arxiv.org/abs/2508.14285