The following papers were recommended by the Semantic Scholar API
\n- \n
- GradientSpace: Unsupervised Data Clustering for Improved Instruction Tuning (2025) \n
- LUNE: Efficient LLM Unlearning via LoRA Fine-Tuning with Negative Examples (2025) \n
- RapidUn: Influence-Driven Parameter Reweighting for Efficient Large Language Model Unlearning (2025) \n
- Fine-Tuning Causal LLMs for Text Classification: Embedding-Based vs. Instruction-Based Approaches (2025) \n
- In-Context Probing for Membership Inference in Fine-Tuned Language Models (2025) \n
- ADAPT: Learning Task Mixtures for Budget-Constrained Instruction Tuning (2025) \n
- Robust LLM-based Column Type Annotation via Prompt Augmentation with LoRA Tuning (2025) \n
Please give a thumbs up to this comment if you found it helpful!
\nIf you want recommendations for any Paper on Hugging Face checkout this Space
\n You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: \n\n@librarian-bot\n\t recommend
Uncertainty-Aware Gradient Signal-to-Noise Data Selection for Instruction Tuning
Abstract
GRADFILTERING is an uncertainty-aware data selection framework for instruction tuning that uses gradient signal-to-noise ratio to improve LLM adaptation efficiency and performance.
Instruction tuning is a standard paradigm for adapting large language models (LLMs), but modern instruction datasets are large, noisy, and redundant, making full-data fine-tuning costly and often unnecessary. Existing data selection methods either build expensive gradient datastores or assign static scores from a weak proxy, largely ignoring evolving uncertainty, and thus missing a key source of LLM interpretability. We propose GRADFILTERING, an objective-agnostic, uncertainty-aware data selection framework that utilizes a small GPT-2 proxy with a LoRA ensemble and aggregates per-example gradients into a Gradient Signal-to-Noise Ratio (G-SNR) utility. Our method matches or surpasses random subsets and strong baselines in most LLM-as-a-judge evaluations as well as in human assessment. Moreover, GRADFILTERING-selected subsets converge faster than competitive filters under the same compute budget, reflecting the benefit of uncertainty-aware scoring.
Community
Instruction tuning is a standard paradigm for adapting large language models (LLMs), but modern instruction datasets are large, noisy, and redundant, making full-data fine-tuning costly and often unnecessary. Existing data selection methods either build expensive gradient datastores or assign static scores from a weak proxy, largely ignoring evolving uncertainty, and thus missing a key source of LLM interpretability. We propose GRADFILTERING, an objective-agnostic, uncertainty-aware data selection framework that utilizes a small GPT-2 proxy with a LoRA ensemble and aggregates per-example gradients into a Gradient Signal-to-Noise Ratio (G-SNR) utility. Our method matches or surpasses random subsets and strong baselines in most LLM-as-a-judge evaluations as well as in human assessment. Moreover, GRADFILTERING-selected subsets converge faster than competitive filters under the same compute budget, reflecting the benefit of uncertainty-aware scoring.
This is an automated message from the Librarian Bot. I found the following papers similar to this paper.
The following papers were recommended by the Semantic Scholar API
- GradientSpace: Unsupervised Data Clustering for Improved Instruction Tuning (2025)
- LUNE: Efficient LLM Unlearning via LoRA Fine-Tuning with Negative Examples (2025)
- RapidUn: Influence-Driven Parameter Reweighting for Efficient Large Language Model Unlearning (2025)
- Fine-Tuning Causal LLMs for Text Classification: Embedding-Based vs. Instruction-Based Approaches (2025)
- In-Context Probing for Membership Inference in Fine-Tuned Language Models (2025)
- ADAPT: Learning Task Mixtures for Budget-Constrained Instruction Tuning (2025)
- Robust LLM-based Column Type Annotation via Prompt Augmentation with LoRA Tuning (2025)
Please give a thumbs up to this comment if you found it helpful!
If you want recommendations for any Paper on Hugging Face checkout this Space
You can directly ask Librarian Bot for paper recommendations by tagging it in a comment:
@librarian-bot
recommend
Models citing this paper 0
No model linking this paper
Datasets citing this paper 0
No dataset linking this paper
Spaces citing this paper 0
No Space linking this paper