Coordinating layer-wise and sentence-wise early exits in LLMs produces multiplicative speedups of 1.4-2.3x over single-dimension early exit on sentiment classification tasks.
hub
DeeBERT: Dynam ic Early Exiting for Accelerating BERT Inference
10 Pith papers cite this work. Polarity classification is still indexing.
10
Pith papers citing it
hub tools
citation-role summary
background 3
citation-polarity summary
roles
background 3polarities
background 3representative citing papers
Training per-layer affine probes on frozen transformers yields more reliable latent predictions than the logit lens and enables detection of malicious inputs from prediction trajectories.
SOL trains a policy to dynamically control multiple efficiency mechanisms per token via group-relative policy optimization on teacher-forced episodes, yielding better quality at matched average budget than static or random allocation.
Looped MoE models scale better than standard transformers because different experts activate on each loop pass, recovering expressivity without extra parameters, and support superior early exits.
Dr. LLM retrofits frozen LLMs with MCTS-supervised per-layer routers for skip/execute/repeat decisions, delivering up to +3.4% accuracy and 5-layer savings on reasoning tasks with strong out-of-domain generalization.
HyperLens reveals that deeper transformer layers magnify small confidence changes into fine-grained trajectories, allowing quantification of cognitive effort where complex tasks demand more and standard SFT can reduce it.
Prerequisite graphs map compressed LLM performance boundaries in analog circuit analysis to allow selecting the smallest viable model for a given task complexity.
Comparative study applies UCB-V, UCB-Tuned, UCB-Bayes and UCB-BwK to ADNN early-exit selection on ResNet and MobileViT using CIFAR-10/100, reporting sub-linear regret with UCB-Bayes fastest and UCB-V/UCB-Tuned best on accuracy-energy and accuracy-latency Pareto fronts.
TOAST approximates full transformer blocks in pretrained models via lightweight closed-form mappings to cut parameters and FLOPs without retraining or finetuning.
A survey synthesizing challenges, system architectures, model optimizations, deployment methods, and resource management techniques for large language model inference at the network edge.
citing papers explorer
-
Sparse Layers are Critical to Scaling Looped Language Models
Looped MoE models scale better than standard transformers because different experts activate on each loop pass, recovering expressivity without extra parameters, and support superior early exits.
-
HyperLens: Quantifying Cognitive Effort in LLMs with Fine-grained Confidence Trajectory
HyperLens reveals that deeper transformer layers magnify small confidence changes into fine-grained trajectories, allowing quantification of cognitive effort where complex tasks demand more and standard SFT can reduce it.
-
Network Edge Inference for Large Language Models: Principles, Techniques, and Opportunities
A survey synthesizing challenges, system architectures, model optimizations, deployment methods, and resource management techniques for large language model inference at the network edge.