Layer by Layer: Uncovering Hidden Representations in Language Models

• URL: http://arxiv.org/abs/2502.02013v2
• Date: Sun, 15 Jun 2025 18:27:17 GMT
• Title: Layer by Layer: Uncovering Hidden Representations in Language Models
• Authors: Oscar Skean, Md Rifat Arefin, Dan Zhao, Niket Patel, Jalal Naghiyev, Yann LeCun, Ravid Shwartz-Ziv,
• Abstract summary: We show that intermediate layers can encode even richer representations, often improving performance on a range of downstream tasks.<n>Our framework highlights how each layer balances information compression and signal preservation, revealing why mid-depth embeddings can exceed the last layer's performance.
• Score: 28.304269706993942
• License: http://creativecommons.org/licenses/by-sa/4.0/
• Abstract: From extracting features to generating text, the outputs of large language models (LLMs) typically rely on the final layers, following the conventional wisdom that earlier layers capture only low-level cues. However, our analysis shows that intermediate layers can encode even richer representations, often improving performance on a range of downstream tasks. To explain and quantify these hidden-layer properties, we propose a unified framework of representation quality metrics based on information theory, geometry, and invariance to input perturbations. Our framework highlights how each layer balances information compression and signal preservation, revealing why mid-depth embeddings can exceed the last layer's performance. Through extensive experiments on 32 text-embedding tasks across various architectures (transformers, state-space models) and domains (language, vision), we demonstrate that intermediate layers consistently provide stronger features, challenging the standard view on final-layer embeddings and opening new directions on using mid-layer representations for more robust and accurate representations.

Related papers

• The Generalization Ridge: Information Flow in Natural Language Generation [7.756342860929851]
  We show how predictive information peaks in upper-middle layers-forming a generalization ridge-before declining in final layers.<n>These findings offer new insights into the internal mechanisms of transformers and underscore the critical role of intermediate layers in supporting generalization.
  arXiv  Detail & Related papers  (2025-07-07T18:18:51Z)
• Intermediate Layer Classifiers for OOD generalization [17.13749013546228]
  In this work, we question the use of last-layer representations for out-of-distribution (OOD) generalisation.<n>We discover that intermediate layer representations frequently offer substantially better generalisation than those from the penultimate layer.<n>Our analysis suggests that intermediate layers are less sensitive to distribution shifts compared to the penultimate layer.
  arXiv  Detail & Related papers  (2025-04-07T19:50:50Z)
• Instruction-Guided Fusion of Multi-Layer Visual Features in Large Vision-Language Models [50.98559225639266]
  We investigate the contributions of visual features from different encoder layers using 18 benchmarks spanning 6 task categories.<n>Our findings reveal that multilayer features provide complementary strengths with varying task dependencies, and uniform fusion leads to suboptimal performance.<n>We propose the instruction-guided vision aggregator, a module that dynamically integrates multi-layer visual features based on textual instructions.
  arXiv  Detail & Related papers  (2024-12-26T05:41:31Z)
• Does Representation Matter? Exploring Intermediate Layers in Large Language Models [22.704926222438456]
  We investigate the quality of intermediate representations in large language models (LLMs)<n>We find that intermediate layers often yield more informative representations for downstream tasks than the final layers.<n>Our results illuminate the internal mechanics of LLMs and guide strategies for architectural optimization and training.
  arXiv  Detail & Related papers  (2024-12-12T18:48:51Z)
• Universal Pooling Method of Multi-layer Features from Pretrained Models for Speaker Verification [7.005068872406135]
  Recent advancements in automatic speaker verification (ASV) studies have been achieved by leveraging large-scale pretrained networks. We present a novel approach for exploiting the multilayered nature of pretrained models for ASV. We show how the proposed interlayer processing aids in maximizing the advantage of utilizing pretrained models.
  arXiv  Detail & Related papers  (2024-09-12T05:55:32Z)
• Entropy Guided Extrapolative Decoding to Improve Factuality in Large Language Models [55.45444773200529]
  Large language models (LLMs) exhibit impressive natural language capabilities but suffer from hallucination. Recent work has focused on decoding techniques to improve factuality during inference.
  arXiv  Detail & Related papers  (2024-04-14T19:45:35Z)
• LayerDiff: Exploring Text-guided Multi-layered Composable Image Synthesis via Layer-Collaborative Diffusion Model [70.14953942532621]
  Layer-collaborative diffusion model, named LayerDiff, is designed for text-guided, multi-layered, composable image synthesis. Our model can generate high-quality multi-layered images with performance comparable to conventional whole-image generation methods. LayerDiff enables a broader range of controllable generative applications, including layer-specific image editing and style transfer.
  arXiv  Detail & Related papers  (2024-03-18T16:28:28Z)
• Exploiting Explainable Metrics for Augmented SGD [43.00691899858408]
  There are several unanswered questions about how learning under optimization really works and why certain strategies are better than others. We propose new explainability metrics that measure the redundant information in a network's layers. We then exploit these metrics to augment the Gradient Descent (SGD) by adaptively adjusting the learning rate in each layer to improve generalization performance.
  arXiv  Detail & Related papers  (2022-03-31T00:16:44Z)
• Feedback Pyramid Attention Networks for Single Image Super-Resolution [37.58180059860872]
  We propose feedback pyramid attention networks (FPAN) to fully exploit the mutual dependencies of features. In our method, the output of each layer in the first stage is also used as the input of the corresponding layer in the next state to re-update the previous low-level filters. We introduce a pyramid non-local structure to model global contextual information in different scales and improve the discriminative representation of the network.
  arXiv  Detail & Related papers  (2021-06-13T11:32:53Z)
• Transformer Feed-Forward Layers Are Key-Value Memories [49.52087581977751]
  We show that feed-forward layers in transformer-based language models operate as key-value memories. We show that the learned patterns are human-interpretable, and that lower layers tend to capture shallow patterns, while upper layers learn more semantic ones.
  arXiv  Detail & Related papers  (2020-12-29T19:12:05Z)
• Understanding and Diagnosing Vulnerability under Adversarial Attacks [62.661498155101654]
  Deep Neural Networks (DNNs) are known to be vulnerable to adversarial attacks. We propose a novel interpretability method, InterpretGAN, to generate explanations for features used for classification in latent variables. We also design the first diagnostic method to quantify the vulnerability contributed by each layer.
  arXiv  Detail & Related papers  (2020-07-17T01:56:28Z)
• Deep Autoencoding Topic Model with Scalable Hybrid Bayesian Inference [55.35176938713946]
  We develop deep autoencoding topic model (DATM) that uses a hierarchy of gamma distributions to construct its multi-stochastic-layer generative network. We propose a Weibull upward-downward variational encoder that deterministically propagates information upward via a deep neural network, followed by a downward generative model. The efficacy and scalability of our models are demonstrated on both unsupervised and supervised learning tasks on big corpora.
  arXiv  Detail & Related papers  (2020-06-15T22:22:56Z)
• Rethinking and Improving Natural Language Generation with Layer-Wise Multi-View Decoding [59.48857453699463]
  In sequence-to-sequence learning, the decoder relies on the attention mechanism to efficiently extract information from the encoder. Recent work has proposed to use representations from different encoder layers for diversified levels of information. We propose layer-wise multi-view decoding, where for each decoder layer, together with the representations from the last encoder layer, which serve as a global view, those from other encoder layers are supplemented for a stereoscopic view of the source sequences.
  arXiv  Detail & Related papers  (2020-05-16T20:00:39Z)
• Global Context-Aware Progressive Aggregation Network for Salient Object Detection [117.943116761278]
  We propose a novel network named GCPANet to integrate low-level appearance features, high-level semantic features, and global context features. We show that the proposed approach outperforms the state-of-the-art methods both quantitatively and qualitatively.
  arXiv  Detail & Related papers  (2020-03-02T04:26:10Z)

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.