SubZero+: Memory-Efficient Adaptive Zeroth-Order LLM Fine-Tuning in Random Subspaces
cs.LG
Submitted: 2026-08-16
Updated: 2026-09-28
Terminology
Sources
- Revisiting the Adam-SGD Gap in LLM Pre-Training: The Role of Large Effective Learning Rates
- The Llama 3 Herd of Models
- Powering Up Zeroth-Order Training via Subspace Gradient Orthogonalization
- The Multi-Query Paradox in Zeroth-Order Optimization
- TeZO: Empowering the Low-Rankness on the Temporal Dimension in the Zeroth-Order Optimization for Fine-tuning LLMs
- Qwen2.5 Technical Report
- SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems
- Qwen3 Technical Report
- OPT: Open Pre-trained Transformer Language Models
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks