Two-dimensional early exit optimisation of LLM inference

• URL: http://arxiv.org/abs/2604.18592v1
• Date: Fri, 27 Mar 2026 15:27:58 GMT
• Title: Two-dimensional early exit optimisation of LLM inference
• Abstract summary: We introduce a two-dimensional (2D) early exit strategy that coordinates layer-wise and sentence-wise exiting for classification tasks in large language models.<n>Our method achieves multiplicative computational savings that exceed those from optimizing either dimension independently.<n>Our findings indicate that 2D early exit strategies excel when semantic information accumulates predictably across input structure.
• Score: 0.4893345190925178
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: We introduce a two-dimensional (2D) early exit strategy that coordinates layer-wise and sentence-wise exiting for classification tasks in large language models. By processing input incrementally sentence-by-sentence while progressively activating deeper layers, our method achieves multiplicative computational savings that exceed those from optimizing either dimension independently. Experimental evaluation across four state-of-the-art LLMs (Llama 3.1, Llama 3.2, Gemma, Qwen; 3B-8B parameters) on three sentiment classification datasets demonstrates additional speed-ups of 1.4--2.3$\times$ over optimal layer-wise early exit for simpler tasks with vanilla models, with graceful degradation on complex multi-class problems. Fine-tuning reduces but does not eliminate this advantage. The approach is model-agnostic, requires only lightweight classification adapters, and is orthogonal to complementary efficiency methods such as quantization and pruning. Our findings indicate that 2D early exit strategies excel when semantic information accumulates predictably across input structure, suggesting possible applicability to sequence-processing tasks beyond sentiment classification.

Related papers

• Prism: Efficient Test-Time Scaling via Hierarchical Search and Self-Verification for Discrete Diffusion Language Models [96.0074341403456]
  Inference-time compute has re-emerged as a practical way to improve LLM reasoning.<n>Most test-time scaling (TTS) algorithms rely on autoregressive decoding.<n>We propose Prism, an efficient TTS framework for dLLMs.
  arXiv  Detail & Related papers  (2026-02-02T09:14:51Z)
• UniC-Lift: Unified 3D Instance Segmentation via Contrastive Learning [6.502142457981839]
  3D Gaussian Splatting (3DGS) and Neural Radiance Fields (NeRF) have advanced novel-view synthesis.<n>Recent methods extend multi-view 2D segmentation to 3D, enabling instance/semantic segmentation for better scene understanding.<n>Key challenge is the inconsistency of 2D instance labels across views, leading to poor 3D predictions.<n>We propose a unified framework that merges these steps, reducing training time and improving performance by introducing a learnable feature embedding for segmentation in Gaussian primitives.
  arXiv  Detail & Related papers  (2025-12-31T10:20:01Z)
• R-Sparse: Rank-Aware Activation Sparsity for Efficient LLM Inference [77.47238561728459]
  R-Sparse is a training-free activation sparsity approach capable of achieving high sparsity levels in advanced LLMs.<n> Experiments on Llama-2/3 and Mistral models across ten diverse tasks demonstrate that R-Sparse achieves comparable performance at 50% model-level sparsity.
  arXiv  Detail & Related papers  (2025-04-28T03:30:32Z)
• Rethinking End-to-End 2D to 3D Scene Segmentation in Gaussian Splatting [86.15347226865826]
  We design a new end-to-end object-aware lifting approach, named Unified-Lift.<n>We augment each Gaussian point with an additional Gaussian-level feature learned using a contrastive loss to encode instance information.<n>We conduct experiments on three benchmarks: LERF-Masked, Replica, and Messy Rooms.
  arXiv  Detail & Related papers  (2025-03-18T08:42:23Z)
• Dynamic layer selection in decoder-only transformers [21.18795712840146]
  We empirically examine two common dynamic inference methods for natural language generation. We find that a pre-trained decoder-only model is significantly more robust to layer removal via layer skipping. We also show that dynamic computation allocation on a per-sequence basis holds promise for significant efficiency gains.
  arXiv  Detail & Related papers  (2024-10-26T00:44:11Z)
• Bypass Back-propagation: Optimization-based Structural Pruning for Large Language Models via Policy Gradient [57.9629676017527]
  We propose an optimization-based structural pruning that learns the pruning masks in a probabilistic space directly by optimizing the loss of the pruned model.<n>We achieve this by learning an underlying Bernoulli distribution to sample binary pruning masks.<n>Experiments conducted on LLaMA, LLaMA-2, LLaMA-3, Vicuna, and Mistral models demonstrate the promising performance of our method in efficiency and effectiveness.
  arXiv  Detail & Related papers  (2024-06-15T09:31:03Z)
• Class-Imbalanced Semi-Supervised Learning for Large-Scale Point Cloud Semantic Segmentation via Decoupling Optimization [64.36097398869774]
  Semi-supervised learning (SSL) has been an active research topic for large-scale 3D scene understanding. The existing SSL-based methods suffer from severe training bias due to class imbalance and long-tail distributions of the point cloud data. We introduce a new decoupling optimization framework, which disentangles feature representation learning and classifier in an alternative optimization manner to shift the bias decision boundary effectively.
  arXiv  Detail & Related papers  (2024-01-13T04:16:40Z)
• Multi-body SE(3) Equivariance for Unsupervised Rigid Segmentation and Motion Estimation [49.56131393810713]
  We present an SE(3) equivariant architecture and a training strategy to tackle this task in an unsupervised manner. Our method excels in both model performance and computational efficiency, with only 0.25M parameters and 0.92G FLOPs.
  arXiv  Detail & Related papers  (2023-06-08T22:55:32Z)
• Unsupervised Learning on 3D Point Clouds by Clustering and Contrasting [11.64827192421785]
  unsupervised representation learning is a promising direction to auto-extract features without human intervention. This paper proposes a general unsupervised approach, named textbfConClu, to perform the learning of point-wise and global features.
  arXiv  Detail & Related papers  (2022-02-05T12:54:17Z)
• DyCo3D: Robust Instance Segmentation of 3D Point Clouds through Dynamic Convolution [136.7261709896713]
  We propose a data-driven approach that generates the appropriate convolution kernels to apply in response to the nature of the instances. The proposed method achieves promising results on both ScanetNetV2 and S3DIS. It also improves inference speed by more than 25% over the current state-of-the-art.
  arXiv  Detail & Related papers  (2020-11-26T14:56:57Z)
• Scan-based Semantic Segmentation of LiDAR Point Clouds: An Experimental Study [2.6205925938720833]
  State of the art methods use deep neural networks to predict semantic classes for each point in a LiDAR scan. A powerful and efficient way to process LiDAR measurements is to use two-dimensional, image-like projections. We demonstrate various techniques to boost the performance and to improve runtime as well as memory constraints.
  arXiv  Detail & Related papers  (2020-04-06T11:08:12Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.