Pith. sign in

REVIEW 21 cited by

From CLIP to DINO: Visual Encoders Shout in Multi-modal Large Language Models

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 2310.08825 v3 pith:QZC3UH5J submitted 2023-10-13 cs.CV

classification cs.CV
keywords visualmllmsclipdinocapabilitiescommencodersfeatures
verification ladder T0 review T1 audit T2 compute T3 formal

Signed reviews

No signed human review yet.

0 comments
read the original abstract

Multi-modal Large Language Models (MLLMs) have made significant strides in expanding the capabilities of Large Language Models (LLMs) through the incorporation of visual perception interfaces. Despite the emergence of exciting applications and the availability of diverse instruction tuning data, existing approaches often rely on CLIP or its variants as the visual branch, and merely extract features from the deep layers. However, these methods lack a comprehensive analysis of the visual encoders in MLLMs. In this paper, we conduct an extensive investigation into the effectiveness of different vision encoders within MLLMs. Our findings reveal that the shallow layer features of CLIP offer particular advantages for fine-grained tasks such as grounding and region understanding. Surprisingly, the vision-only model DINO, which is not pretrained with text-image alignment, demonstrates promising performance as a visual branch within MLLMs. By simply equipping it with an MLP layer for alignment, DINO surpasses CLIP in fine-grained related perception tasks. Building upon these observations, we propose a simple yet effective feature merging strategy, named COMM, that integrates CLIP and DINO with Multi-level features Merging, to enhance the visual capabilities of MLLMs. We evaluate COMM through comprehensive experiments on a wide range of benchmarks, including image captioning, visual question answering, visual grounding, and object hallucination. Experimental results demonstrate the superior performance of COMM compared to existing methods, showcasing its enhanced visual capabilities within MLLMs.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 21 Pith papers

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

  1. Mitigating Visual Degradation in MLLMs via Spatial-Spectral Visual Anchor Learning

    cs.CV 2026-08 conditional novelty 6.0 of 10

    SSVAL uses visual anchor prompts plus spatial and frequency-domain alignment losses to reduce visual information loss in MLLMs and improve fine-grained perception benchmarks.

  2. InCoM: Intent-Driven Perception and Structured Coordination for Mobile Manipulation

    cs.RO 2026-02 unverdicted novelty 6.0 of 10

    InCoM reports 23–28 percentage-point success-rate gains in mobile manipulation benchmarks by dynamically reweighting multi-scale perception via inferred motion intent and decoupling base-arm action generation with flo...

  3. Less is Enough: Synthesizing Diverse Data in LLM Feature Space with Sparse Autoencoders

    cs.CL 2026-02 conditional novelty 6.0 of 10

    Coverage of sparse-autoencoder-identified task features predicts post-training performance and can guide synthesis of small, high-impact datasets (2,000 vs. 300,000 samples).

  4. Harmonia: Algorithm-Hardware Co-Design for Memory- and Compute-Efficient BFP-based LLM Inference

    cs.AR 2026-02 conditional novelty 6.0 of 10

    Harmonia runs LLM inference with all activations in block floating point (BFP) and a 4-bit KV cache, reporting 3.08x average speedup, 2.03x energy savings, and under 1% accuracy loss on LongBench.

  5. Who Endorsed It? Measuring Authority Bias Across Expertise Levels in Language Models

    cs.CL 2026-01 reject novelty 6.0 of 10

    Language models accept misleading endorsements more readily from higher-expertise persona sources, and the effect can be partially steered away by subtracting a residual-stream vector.

  6. VKnowU: Evaluating Visual Knowledge Understanding in Multimodal LLMs

    cs.CV 2025-11 conditional novelty 6.0 of 10

    A 1,680-question video benchmark shows leading multimodal models lag humans by ~15 points on visual knowledge, and a See-Think-Answer RL-trained model narrows the gap.

  7. Testing for LLM response differences: the case of a composite null consisting of semantically irrelevant query perturbations

    math.ST 2025-09 conditional novelty 6.0 of 10

    A new hypothesis test for binary LLM responses treats semantically equivalent query perturbations as an unknown null set and gives asymptotic validity and consistency guarantees under a uniformity assumption.

  8. TMUAD: Enhancing Logical Capabilities in Unified Anomaly Detection Models with a Text Memory Bank

    cs.CV 2025-08 conditional novelty 6.0 of 10

    TMUAD detects both structural and logical anomalies by comparing a query image's textual object descriptions, object crops, and image patches against three normal memory banks, reaching state-of-the-art AUROC on seven...

  9. Decoding Memories: An Efficient Pipeline for Self-Consistency Hallucination Detection

    cs.CL 2025-08 conditional novelty 6.0 of 10

    A decoding pipeline reuses cached tokens and anneals sampling temperature to accelerate self-consistency hallucination detection by up to 3x without meaningful AUROC loss.

  10. CompressKV: Semantic Retrieval Heads Know What Tokens are Not Important Before Generation

    cs.CL 2025-08 unverdicted novelty 6.0 of 10

    A method that uses a handful of 'semantic retrieval heads' instead of all attention heads to decide which key-value cache entries can be dropped, plus layer-wise cache budgeting, reportedly beats prior KV compression ...

  11. METEOR: Multi-Encoder Collaborative Token Pruning for Efficient Vision Language Models

    cs.CV 2025-07 conditional novelty 6.0 of 10

    METEOR is a three-stage token pruning framework that reduces visual tokens in multi-encoder MLLMs by 76% with only a 0.3% average accuracy drop.

  12. Towards Multimodal Understanding via Stable Diffusion as a Task-Aware Feature Extractor

    cs.CV 2025-07 conditional novelty 6.0 of 10

    Stable Diffusion features, especially when conditioned on the question, improve vision-centric multimodal question answering when fused with CLIP.

  13. Kernel-based Unsupervised Embedding Alignment for Enhanced Visual Representation in Vision-language Models

    cs.CV 2025-06 conditional novelty 6.0 of 10

    Fine-tuning CLIP's visual encoder to match DINOv2's kernel-based similarity structure improves its fine-grained visual perception while preserving its alignment to text.

  14. Diffusion Instruction Tuning

    cs.LG 2025-02 conditional novelty 6.0 of 10

    Lavender fine-tunes vision-language models by aligning their attention maps with Stable Diffusion's attention targets, improving accuracy on 20 benchmarks with as few as 0.13 million training examples.

  15. CPath-Omni: A Unified Multimodal Foundation Model for Patch and Whole Slide Image Analysis in Computational Pathology

    cs.CV 2024-12 conditional novelty 6.0 of 10

    A single 15B pathology model unifies patch-level and whole-slide tasks and reports state-of-the-art results on 39 of 42 benchmark datasets.

  16. Libra: Leveraging Temporal Images for Biomedical Radiology Analysis

    cs.CV 2024-11 conditional novelty 6.0 of 10

    Libra introduces a Temporal Alignment Connector for multimodal LLMs that fuses current and prior chest X-ray features and reports improved radiology report generation on MIMIC-CXR.

  17. Advancing Fine-Grained Visual Understanding with Multi-Scale Alignment in Multi-Modal Models

    cs.CV 2024-11 conditional novelty 6.0 of 10

    A three-stage training method that aligns object texts, coordinates, and cropped images, plus a GPT-4V-based data synthesis pipeline, lets 1.5B and 3B models match or exceed larger MLLMs on grounding and VQA benchmarks.

  18. Interpretable Face Anti-Spoofing: Enhancing Generalization with Multimodal Large Language Models

    cs.CV 2025-01 conditional novelty 5.0 of 10

    I-FAS reformulates face anti-spoofing as interpretable VQA with spoof-aware captions, a lopsided language-model loss, and a globally aware connector, achieving state-of-the-art cross-domain results.

  19. ComprehendEdit: A Comprehensive Dataset and Evaluation Framework for Multimodal Knowledge Editing

    cs.CV 2024-12 conditional novelty 5.0 of 10

    A new eight-task benchmark with in-domain metrics KGI and KPI reveals that existing multimodal editing methods degrade on related samples, and the proposed HICE method achieves a better balance.

  20. VidHalluc: Evaluating Temporal Hallucinations in Multimodal Large Language Models for Video Understanding

    cs.CV 2024-12 conditional novelty 5.0 of 10

    VidHalluc is a 5,002-video paired benchmark for action, temporal sequence, and scene transition hallucinations in video MLLMs, and DINO-HEAL is a training-free saliency reweighting method that improves hallucination s...

  21. HEAL: A Hypothesis-Based Preference-Aware Analysis Framework

    cs.CL 2025-08 conditional novelty 4.0 of 10

    HEAL evaluates preference optimization by measuring ranking accuracy and strength correlation between model likelihoods and proxy reward scores over multi-response hypothesis spaces.

Pith tools