Pith. sign in

REVIEW 2 cited by

Frozen Transformers in Language Models Are Effective Visual Encoder Layers

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2310.12973 v2 pith:5JKIOPCS submitted 2023-10-19 cs.CV cs.AIcs.CLcs.LG

classification cs.CVcs.AIcs.CLcs.LG
keywords tasksvisualllmslanguagetransformerblockspre-trainedencoder
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

This paper reveals that large language models (LLMs), despite being trained solely on textual data, are surprisingly strong encoders for purely visual tasks in the absence of language. Even more intriguingly, this can be achieved by a simple yet previously overlooked strategy -- employing a frozen transformer block from pre-trained LLMs as a constituent encoder layer to directly process visual tokens. Our work pushes the boundaries of leveraging LLMs for computer vision tasks, significantly departing from conventional practices that typically necessitate a multi-modal vision-language setup with associated language prompts, inputs, or outputs. We demonstrate that our approach consistently enhances performance across a diverse range of tasks, encompassing pure 2D and 3D visual recognition tasks (e.g., image and point cloud classification), temporal modeling tasks (e.g., action recognition), non-semantic tasks (e.g., motion forecasting), and multi-modal tasks (e.g., 2D/3D visual question answering and image-text retrieval). Such improvements are a general phenomenon, applicable to various types of LLMs (e.g., LLaMA and OPT) and different LLM transformer blocks. We additionally propose the information filtering hypothesis to explain the effectiveness of pre-trained LLMs in visual encoding -- the pre-trained LLM transformer blocks discern informative visual tokens and further amplify their effect. This hypothesis is empirically supported by the observation that the feature activation, after training with LLM transformer blocks, exhibits a stronger focus on relevant regions. We hope that our work inspires new perspectives on utilizing LLMs and deepening our understanding of their underlying mechanisms. Code is available at https://github.com/ziqipang/LM4VisualEncoding.

Discussion (0). Sign in to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Language-Unlocked ViT (LUViT): Empowering Self-Supervised Vision Transformers with LLMs

    cs.CV 2025-07 conditional novelty 6.0 of 10

    LUViT jointly pretrains a ViT with masked auto-encoding and LoRA adapters in a frozen LLM block, reporting +0.4% ImageNet-1K accuracy and up to +2.2% on ImageNet-A over its own MAE baseline.

  2. Pre-Trained LLM is a Semantic-Aware and Generalizable Segmentation Booster

    cs.CV 2025-06 conditional novelty 6.0 of 10

    A frozen pre-trained LLM layer inserted between a CNN encoder and decoder improves medical image segmentation across ultrasound, dermoscopy, polyp, and CT benchmarks with few added trainable parameters.

Pith tools