Rethinking Data Selection: The Importance of Coverage over Difficulty in LLM Fine-Tuning
cs.LG
Submitted: 2026-11-05
Updated: 2026-11-05
Code: https://github.com/LalchandPandia/RethinkingDataSelection
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- SemDeDup: Data-efficient learning at web-scale through semantic deduplication
- Estimating Example Difficulty Using Variance of Gradients
- Perplexed by Perplexity: Perplexity-Based Data Pruning With Small Reference Models
- Deep Batch Active Learning by Diverse, Uncertain Gradient Lower Bounds
- Depthwise Hyperparameter Transfer in Residual Networks: Dynamics and Scaling Limit
- Evaluating Large Language Models Trained on Code
- Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge
- The Llama 3 Herd of Models
- Measuring Massive Multitask Language Understanding
- Training Compute-Optimal Large Language Models
- Scaling Laws for Neural Language Models
- Tulu 3: Pushing Frontiers in Open Language Model Post-Training
- D2 Pruning: Message Passing for Balancing Diversity and Difficulty in Data Pruning
- When Less is More: Investigating Data Pruning for Pretraining LLMs at Scale
- 2 OLMo 2 Furious
- How to Train Data-Efficient LLMs
- LIMO: Less is More for Reasoning
- Coverage-centric Coreset Selection for High Pruning Rates
Related papers
- A Physics-Regulated Neural Framework for Learning 3D Grain Growth Dynamics
- FlexMoE: Learning One-for-All Multi-Budget Expert Pruning for Flexible MoE Deployment
- OptiMind: Improving Small Language Models for Optimization Modeling through Formulation Guidance and Multi-Turn Learning
- Staying in the Sweet Spot: Boost Reasoning Evolution via Responsive Rollout Difficulty Manipulation
- Examining the Impact of Data and Training Methods on Emergent and Subliminal Misalignment
- Rethinking Black-Box On-Policy Distillation: A Rubric-based Analysis