Astra: Activation-Space Tail-Eigenvector Low-Rank Adaptation of Large Language Models
This work addresses a bottleneck in efficient adaptation of large language models for downstream tasks, offering an incremental improvement over existing PEFT methods like LoRA.
The paper tackled the problem of suboptimal fine-tuning performance in Parameter-Efficient Fine-Tuning (PEFT) methods by proposing Astra, a novel method that leverages tail eigenvectors of model output activations to construct task-adaptive low-rank adapters, resulting in faster convergence and improved performance across 16 benchmarks, even surpassing full fine-tuning in some cases.
Parameter-Efficient Fine-Tuning (PEFT) methods, especially LoRA, are widely used for adapting pre-trained models to downstream tasks due to their computational and storage efficiency. However, in the context of LoRA and its variants, the potential of activation subspaces corresponding to tail eigenvectors remains substantially under-exploited, which may lead to suboptimal fine-tuning performance. In this work, we propose Astra (Activation-Space Tail-Eigenvector Low-Rank Adaptation), a novel PEFT method that leverages the tail eigenvectors of the model output activations-estimated from a small task-specific calibration set-to construct task-adaptive low-rank adapters. By constraining updates to the subspace spanned by these tail eigenvectors, Astra achieves faster convergence and improved downstream performance with a significantly reduced parameter budget. Extensive experiments across natural language understanding (NLU) and natural language generation (NLG) tasks demonstrate that Astra consistently outperforms existing PEFT baselines across 16 benchmarks and even surpasses full fine-tuning (FFT) in certain scenarios.