FLAMIA: FLAsh MIcroscaling Attention for Fast and Causally Safe Training on Blackwell Paper • 2611.03344 • Published 16 days ago • 22
Policy-Mask: Budget-Aware Adaptation of Diffusion Features to Discriminative Tasks Paper • 2610.28889 • Published 30 days ago • 48
Kandinsky 6.0 Video: Foundation Models for Synchronized Video and Audio Generation Paper • 2610.05608 • Published Oct 4 • 173
LANTERN: Illuminating Hidden Mathematical Knowledge in Language Models Paper • 2609.32264 • Published Sep 26 • 34
Disaggregated Quantization: Specializing LLM Prefill and Decode Paper • 2609.26333 • Published Sep 22 • 83
Your Transformer Can Hold Two Thoughts at Once: Evidence of Linear Superposition in LLMs Paper • 2609.29845 • Published Sep 24 • 90
Ouroboros: A Self-Developing Frontier Coding Agent with Reviewed Core Evolution Paper • 2608.08311 • Published Aug 8 • 92
One-Step Gradient Delay is Not a Barrier for Large-Scale Asynchronous Pipeline Parallel LLM Pretraining Paper • 2606.30634 • Published Jun 29 • 23
MatryoshkaLoRA: Learning Accurate Hierarchical Low-Rank Representations for LLM Fine-Tuning Paper • 2605.07850 • Published May 8 • 18
Reasoning Shift: How Context Silently Shortens LLM Reasoning Paper • 2604.01161 • Published Apr 1 • 31
view article Article Unlocking Agentic RL Training for GPT-OSS: A Practical Retrospective LinkedIn • Jan 27 • 82
LK Losses: Direct Acceptance Rate Optimization for Speculative Decoding Paper • 2602.23881 • Published Feb 27 • 18
MatGPTQ: Accurate and Efficient Post-Training Matryoshka Quantization Paper • 2602.03537 • Published Feb 3 • 7
DASH: Faster Shampoo via Batched Block Preconditioning and Efficient Inverse-Root Solvers Paper • 2602.02016 • Published Feb 2 • 15
Quartet II: Accurate LLM Pre-Training in NVFP4 by Improved Unbiased Gradient Estimation Paper • 2601.22813 • Published Jan 30 • 63
WUSH: Near-Optimal Adaptive Transforms for LLM Quantization Paper • 2512.00956 • Published Nov 30, 2025 • 25
view article Article A Review on the Evolvement of Load Balancing Strategy in MoE LLMs: Pitfalls and Lessons NormalUhr • Feb 4, 2025 • 38
view article Article Introducing Trackio: A Lightweight Experiment Tracking Library from Hugging Face +3 abidlabs, znation, nouamanetazi, sasha, qgallouedec • Jul 29, 2025 • 232