What the Loss Curve Misses: FP4 Pretraining Slows the Acquisition of Rare Facts
cs.LG
Submitted: 2026-11-09
Updated: 2026-11-09
Terminology
Sources
- What Matters for NVFP4 Training? A Comparative Study of Low-Precision Pretraining Recipes
- Eliciting Latent Predictions from Transformers with the Tuned Lens
- The Curse and Blessing of Mean Bias in FP4-Quantized LLM Training
- DeepSeek-V3 Technical Report
- Muon is Scalable for LLM Training
- An Empirical Model of Large-Batch Training
- FP8 Formats for Deep Learning
- Pretraining Large Language Models with NVFP4
- Linger and Lose: Knowledge Collapse in Low-Bit Language Models
- FP8-LM: Training FP8 Large Language Models
- Microscaling Data Formats for Deep Learning
- GLU Variants Improve Transformer
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