The Structure of Quantization Damage in LLMs: Why the Next Bit Should Be Spent Globally
cs.LG, cs.CL
Submitted: 2026-09-01
Updated: 2026-09-01
Comments: Preprint. Under review at a NeurIPS 2026 workshop. 11 pages, 4 figures, 8 tables
Code: https://github.com/meta-llama/llama3
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- GPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformers
- Mistral 7B
- You Had One Job: Per-Task Quantization Using LLMs' Hidden Representations
- Qwen3 Technical Report
- Task-Circuit Quantization: Leveraging Knowledge Localization and Interpretability for Compression
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: Local-Consistency Optimisation for Bayesian Neural Networks
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks