Low-precision Flash Attention fails due to similar low-rank attention representations combined with biased rounding errors that accumulate and corrupt weight updates; a minimal fix to reduce rounding bias stabilizes training.
A theory on Adam instability in large-scale machine learning.arXiv preprint arXiv:2304.09871,
9 Pith papers cite this work. Polarity classification is still indexing.
9
Pith papers citing it
citation-role summary
background 3
method 1
citation-polarity summary
representative citing papers
MiniMax-M1 is a 456B parameter hybrid-attention MoE model trained with CISPO RL that achieves performance comparable or superior to DeepSeek-R1 and Qwen3-235B on reasoning and software engineering tasks while training in three weeks on 512 GPUs.
GWT projects gradients into wavelet subspaces to compress optimizer states for memory-efficient LLM training while claiming performance parity with full-rank updates.
BAGEL is a unified decoder-only model that develops emerging complex multimodal reasoning abilities after pretraining on large-scale interleaved data and outperforms prior open-source unified models.
Open-Sora releases an open-source video generation model based on a Spatial-Temporal Diffusion Transformer that decouples spatial and temporal attention, supporting text-to-video, image-to-video, and text-to-image tasks with claimed high fidelity.
PowLU replaces SwiGLU with a rational-power activation to reduce outlier amplification and numerical instability during large-scale LLM pre-training while matching performance.
Gemini 2.5 Pro and Flash models are presented as achieving frontier performance in reasoning, coding, and long-context multimodal tasks while spanning a cost-capability Pareto curve.
This survey organizes LLM optimizer literature into categories and argues the field is shifting toward rigorous, multi-factor comparisons of convergence, memory, stability, and complexity.
citing papers explorer
-
Why Low-Precision Transformer Training Fails: An Analysis on Flash Attention
Low-precision Flash Attention fails due to similar low-rank attention representations combined with biased rounding errors that accumulate and corrupt weight updates; a minimal fix to reduce rounding bias stabilizes training.
-
MiniMax-M1: Scaling Test-Time Compute Efficiently with Lightning Attention
MiniMax-M1 is a 456B parameter hybrid-attention MoE model trained with CISPO RL that achieves performance comparable or superior to DeepSeek-R1 and Qwen3-235B on reasoning and software engineering tasks while training in three weeks on 512 GPUs.
-
GWT: Scalable Optimizer State Compression for Large Language Model Training
GWT projects gradients into wavelet subspaces to compress optimizer states for memory-efficient LLM training while claiming performance parity with full-rank updates.
-
Emerging Properties in Unified Multimodal Pretraining
BAGEL is a unified decoder-only model that develops emerging complex multimodal reasoning abilities after pretraining on large-scale interleaved data and outperforms prior open-source unified models.
-
Open-Sora: Democratizing Efficient Video Production for All
Open-Sora releases an open-source video generation model based on a Spatial-Temporal Diffusion Transformer that decouples spatial and temporal attention, supporting text-to-video, image-to-video, and text-to-image tasks with claimed high fidelity.
-
PowLU: An Activation Function for Stable Pre-Training of LLMs
PowLU replaces SwiGLU with a rational-power activation to reduce outlier amplification and numerical instability during large-scale LLM pre-training while matching performance.
-
Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities
Gemini 2.5 Pro and Flash models are presented as achieving frontier performance in reasoning, coding, and long-context multimodal tasks while spanning a cost-capability Pareto curve.
-
Navigating LLM Valley: From AdamW to Memory-Efficient and Matrix-Based Optimizers
This survey organizes LLM optimizer literature into categories and argues the field is shifting toward rigorous, multi-factor comparisons of convergence, memory, stability, and complexity.