# Which and When to Admit: Gradient Admission for Data-Centric Small Language Model Finetuning

arXiv: https://arxiv.org/abs/2610.07553 · HTML: https://mlsanity.com/paper/2610.07553/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 29 of its 38 references that are in the map.
- First posted on arXiv 6 Oct 2026.

## Builds on (most cited first, 29 of 29)

- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [Understanding Black-box Predictions via Influence Functions](https://mlsanity.com/paper/1703.04730/) (2017)
- [Gradient Surgery for Multi-Task Learning](https://mlsanity.com/paper/2001.06782/) (2020)
- [DoRA: Weight-Decomposed Low-Rank Adaptation](https://mlsanity.com/paper/2402.09353/) (2024)
- [An Empirical Study of Example Forgetting during Deep Neural Network Learning](https://mlsanity.com/paper/1812.05159/) (2018)
- [Intrinsic Dimensionality Explains the Effectiveness of Language Model Fine-Tuning](https://mlsanity.com/paper/2012.13255/) (2020)
- [Estimating Training Data Influence by Tracing Gradient Descent](https://mlsanity.com/paper/2002.08484/) (2020)
- [Deep Learning on a Data Diet: Finding Important Examples Early in Training](https://mlsanity.com/paper/2107.07075/) (2021)
- [LESS: Selecting Influential Data for Targeted Instruction Tuning](https://mlsanity.com/paper/2402.04333/) (2024)
- [GaLore: Memory-Efficient LLM Training by Gradient Low-Rank Projection](https://mlsanity.com/paper/2403.03507/) (2024)
- [Conflict-Averse Gradient Descent for Multi-task Learning](https://mlsanity.com/paper/2110.14048/) (2021)
- [LoRA+: Efficient Low Rank Adaptation of Large Models](https://mlsanity.com/paper/2402.12354/) (2024)
- [AdaLoRA: Adaptive Budget Allocation for Parameter-Efficient Fine-Tuning](https://mlsanity.com/paper/2303.10512/) (2023)
- [Fine-Tuning Language Models with Just Forward Passes](https://mlsanity.com/paper/2305.17333/) (2023)
- [GRAD-MATCH: Gradient Matching based Data Subset Selection for Efficient Deep Model Training](https://mlsanity.com/paper/2103.00123/) (2021)
- [PiSSA: Principal Singular Values and Singular Vectors Adaptation of Large Language Models](https://mlsanity.com/paper/2404.02948/) (2024)
- [LoRA-Pro: Are Low-Rank Adapters Properly Optimized?](https://mlsanity.com/paper/2407.18242/) (2024)
- [Variance-reduced Zeroth-Order Methods for Fine-Tuning Language Models](https://mlsanity.com/paper/2404.08080/) (2024)
- [X-LoRA: Mixture of Low-Rank Adapter Experts, a Flexible Framework for Large Language Models with Applications in Protein Mechanics and Molecular Design](https://mlsanity.com/paper/2402.07148/) (2024)
- [G-DIG: Towards Gradient-based Diverse and High-quality Instruction Data Selection for Machine Translation](https://mlsanity.com/paper/2405.12915/) (2024)
- [Zeroth-Order Fine-Tuning of LLMs in Random Subspaces](https://mlsanity.com/paper/2410.08989/) (2024)
- [ConPET: Continual Parameter-Efficient Tuning for Large Language Models](https://mlsanity.com/paper/2309.14763/) (2023)
- [Continuous Optimization for Feature Selection with Permutation-Invariant Embedding and Policy-Guided Search](https://mlsanity.com/paper/2505.11601/) (2025)
- [CTR-LoRA: Curvature-Aware and Trust-Region Guided Low-Rank Adaptation for Large Language Models](https://mlsanity.com/paper/2510.15962/) (2025)
- [FLOPS: Forward Learning with OPtimal Sampling](https://mlsanity.com/paper/2410.05966/) (2024)
- [MoA: Heterogeneous Mixture of Adapters for Parameter-Efficient Fine-Tuning of Large Language Models](https://mlsanity.com/paper/2506.05928/) (2025)
- [Sim2Act: Robust Simulation-to-Decision Learning via Adversarial Calibration and Group-Relative Perturbation](https://mlsanity.com/paper/2603.09053/) (2026)
- [Mitigating Shortcut Reasoning in Language Models: A Gradient-Aware Training Approach](https://mlsanity.com/paper/2603.20899/) (2026)
- [Rethinking Data Augmentation under Covariate Shift: Invariant-Guided Diffusion and Prototype Reweighting](https://mlsanity.com/paper/2610.00873/) (2026)

## Abstract (arXiv)

LoRA fine-tuning adapts small language models (SLMs) to heterogeneous instruction data within a low-rank update subspace, making it vulnerable to three structural problems: conflicting gradients that cancel, static data selection that cannot track evolving learning dynamics, and subspace saturation that causes later updates to overwrite useful directions. We argue that effective adaptation therefore requires controlling which data-induced gradients enter the LoRA subspace and when. We propose GRADE (GRadient-Aligned Data-centric rEcipe), a data-centric framework combining two mechanisms: a state-aware selector that continually admits samples aligned with the evolving multi-task gradient field, and a self-calibrating step-level gate that rejects updates likely to cause destructive overwrite near saturation. Across three current-generation backbones and a heterogeneous seven-dataset instruction pool, GRADE outperforms strong data-selection and PEFT-stabilization baselines in accuracy and robustness. It is the only method to improve consistently over standard LoRA on every architecture, while producing more coherent gradient trajectories and less destructive overwrite. These results show that successful SLM adaptation depends not only on which data are selected, but also on which gradients are allowed to enter and persist in the constrained update subspace.