arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2605.12207cs.LGcs.AIcs.CL

不是数量而是选择:低秩适应中的参数放置

Not How Many, But Which: Parameter Placement in Low-Rank Adaptation

Kensho Technologies

Arijit Sehanobish, Charles Lovering

AI总结:

研究参数放置对LoRA适配器性能的影响,发现随机与有信息的子集在监督微调中表现相当,而在基模型上GRPO时,有信息放置能恢复标准LoRA精度,归因于梯度结构差异。

摘要:

我们研究参数放置问题:给定LoRA适配器B矩阵中固定预算k的可训练条目,选择哪些k条目是否重要?在监督微调中,随机和有信息的子集表现相当。在基模型上的GRPO中,随机放置无法提升基模型性能,而梯度有信息放置能恢复标准LoRA精度。这种制度依赖性源于梯度结构:SFT梯度低秩且方向稳定,因此任何子集都能积累一致更新;GRPO梯度高秩且跨步近正交,因此只有具有一致符号梯度的元素能保留学习信号。我们的评分过程在不到10秒内以低于0.5%的训练成本识别出这些关键参数。所选参数集中在残差流写入投影(V,O,Down),在不同模型家族和规模(1.5B-8B)中稳定。

英文摘要:

We study the \textit{parameter placement problem}: given a fixed budget of $k$ trainable entries within the B matrix of a LoRA adapter (A frozen), does the choice of which $k$ matter? Under supervised fine-tuning, random and informed subsets achieve comparable performance. Under GRPO on base models, random placement fails to improve over the base model, while gradient-informed placement recovers standard LoRA accuracy. This regime dependence traces to gradient structure: SFT gradients are low-rank and directionally stable, so any subset accumulates coherent updates; GRPO gradients are high-rank and near-orthogonal across steps, so only elements with consistently signed gradients retain the learning signal. Our scoring procedure identifies these critical parameters in under 10 seconds at less than 0.5% of training cost. Selected parameters concentrate on residual-stream-writing projections (V, O, Down), stable across model families and scales (1.5B - 8B).

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑