repoGitLabTrust 82 · PrimaryPublished 3d agoLive · 3d ago
zwmaronek/Beyond-Early-Exit
Beyond Early Exit: Solving GPU Warp Divergence in Adaptive LLM Inference with Micro-Batched Routing. Author: Zachary Maronek Date: January 2026
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Why these links exist
Every edge carries a method, confidence, and the source snippet that justified it — so bad links are debuggable.
- PossiblePossibly related (embedding) · 56%OpenAI and Broadcom announce chip designed for LLM inference at scale →
- PossiblePossibly related (embedding) · 56%WattGPU: Predicting Inference Power and Latency on Unseen GPUs and LLMs →
- PossiblePossibly related (embedding) · 53%Depth Exploration for LLM Decoding →
- PossiblePossibly related (embedding) · 53%Evaluate a model properly →
- PossiblePossibly related (embedding) · 52%Efficient PEFT Methods with Adaptive Checkpointing for Vision Models and VLMs on Resource Constrained Consumer-GPUs →
- PossiblePossibly related (embedding) · 48%The Geometry of Memorization: Finite-Time Spectral Sensitivity as a Diagnostic for Flow Matching Models →
- PossiblePossibly related (embedding) · 53%Accelerating Block Low-Rank Foundation Model Inference on MemoryConstrained GPUs →
Covers
Implements
Related to
Implements (incoming)
Covers (incoming)
Related across the graph
paperEfficient PEFT Methods with Adaptive Checkpointing for Vision Models and VLMs on Resource Constrained Consumer-GPUsnewsOpenAI and Broadcom announce chip designed for LLM inference at scalenewsAccelerating Block Low-Rank Foundation Model Inference on MemoryConstrained GPUspaperWattGPU: Predicting Inference Power and Latency on Unseen GPUs and LLMspaperThe Geometry of Memorization: Finite-Time Spectral Sensitivity as a Diagnostic for Flow Matching ModelspaperDepth Exploration for LLM DecodingtutorialEvaluate a model properly