Nexus: Same Pretraining Loss, Better Downstream Generalization via Common Minima
cs.LG
Submitted: 2026-04-10
Updated: 2026-11-05
Code: https://github.com/ByteDance-Seed/seed-oss
Terminology
Sources
- Program Synthesis with Large Language Models
- NVIDIA Nemotron Nano 2: An Accurate and Efficient Hybrid Mamba-Transformer Reasoning Model
- Unveiling the Basin-Like Loss Landscape in Large Language Models
- Training Verifiers to Solve Math Word Problems
- Gradient Descent on Neural Networks Typically Occurs at the Edge of Stability
- Self-Stabilization: The Implicit Bias of Gradient Descent at the Edge of Stability
- The Llama 3 Herd of Models
- Evaluating Large Language Models Trained on Code
- Sharpness-Aware Minimization for Efficiently Improving Generalization
- Gradient Descent Happens in a Tiny Subspace
- HyperNetworks
- Measuring Massive Multitask Language Understanding
- Measuring Mathematical Problem Solving With the MATH Dataset
- MiniCPM: Unveiling the Potential of Small Language Models with Scalable Training Strategies
- Three Factors Influencing Minima in SGD
- Scaling Laws for Neural Language Models
- On Large-Batch Training for Deep Learning: Generalization Gap and Sharp Minima
- Pre-training under infinite compute
- Adam: A Method for Stochastic Optimization
- Meta-SGD: Learning to Learn Quickly for Few-Shot Learning
Related papers
- 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
- (Mis)generalization of Helpful-Only Fine-tuning