Paper Citation Record · LEDGER

An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

As of 12 August 2026, this Paper Citation Record lists 0 of 0 outbound references and 36 inbound Pith citation observations for arXiv:2403.06764.

A citation records a reference. It does not transfer a finding from one paper to another.

pith.paper-citation-record.v1
2403.06764 v3

Coverage vector

measured 0 of 0 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links

measured 36 of 36 standing notices

One-hop event checks from named stored sources.

Source: scholarly_work_events, retraction_status_cache, observed 2026-08-11T06:34:44.6726+00:00

measured 36 of 36 inbound itemization

Pith citing papers itemized under the disclosed page cap.

Source: paper_references, paper_reference_links, observed 2026-08-12T00:57:32.397998Z

measured 1 of 1 external citation measurements

A source-named dated measurement, never combined with another source.

Source: pith, observed 2026-08-05T02:28:24.338817Z

Reference resolution

0 of 0 outbound references displayed

verified exact0
verified fuzzy0
unresolved0
parse uncertain0
malformed identifier0
metadata mismatch0

External citation measurements

1
pith, observed 2026-08-05T02:28:24.338817Z

Outbound references

No outbound reference observations are available for this paper version.

Pith citing papers

Observation 36fb42a2-82c2-44de-8e31-e2539c974c31 · inbound

PyramidKV: Dynamic KV Cache Compression based on Pyramidal Information Funneling cites this paper.

PyramidKV: Dynamic KV Cache Compression based on Pyramidal Information Funneling An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 3

Resolution: verified exact; arxiv_id, observed 2026-05-12T09:58:29.106770Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation 9129ef6c-85f5-4afb-9e3d-acebd4def756 · inbound

When Attention Sink Emerges in Language Models: An Empirical View cites this paper.

When Attention Sink Emerges in Language Models: An Empirical View An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 6

Resolution: verified exact; arxiv_id, observed 2026-05-16T17:41:03.797181Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation bd0e560b-a626-4864-acc2-3525f206e579 · inbound

PyramidDrop: Accelerating Your Large Vision-Language Models via Pyramid Visual Redundancy Reduction cites this paper.

PyramidDrop: Accelerating Your Large Vision-Language Models via Pyramid Visual Redundancy Reduction An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 9

Resolution: verified exact; arxiv_id, observed 2026-05-15T12:12:14.760555Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation 2f00ea8d-456e-46c8-bfde-d319e9726ef2 · inbound

Beyond Text-Visual Attention: Exploiting Visual Cues for Effective Token Pruning in VLMs cites this paper.

Beyond Text-Visual Attention: Exploiting Visual Cues for Effective Token Pruning in VLMs An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 9

Resolution: unresolved; no resolver link, observed 2026-08-12T00:57:32.397998Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9ece6cac-4691-4b1a-8bbb-3c84f29e0dc9 · inbound

Explainable and Interpretable Multimodal Large Language Models: A Comprehensive Survey cites this paper.

Explainable and Interpretable Multimodal Large Language Models: A Comprehensive Survey An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 136

Resolution: unresolved; no resolver link, observed 2026-08-11T23:54:23.806149Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7a680d99-9396-4a95-a4fa-dd313db57ba8 · inbound

A Stitch in Time Saves Nine: Small VLM is a Precise Guidance for Accelerating Large VLMs cites this paper.

A Stitch in Time Saves Nine: Small VLM is a Precise Guidance for Accelerating Large VLMs An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 5

Resolution: unresolved; no resolver link, observed 2026-08-11T22:35:24.150592Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c5ce3a98-1b80-4e69-8f83-431aa4261138 · inbound

FlashSloth: Lightning Multimodal Large Language Models via Embedded Visual Compression cites this paper.

FlashSloth: Lightning Multimodal Large Language Models via Embedded Visual Compression An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 7

Resolution: unresolved; no resolver link, observed 2026-08-11T21:37:08.246133Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a1f51f80-4cd2-4477-b5aa-eaf8a52d8ad6 · inbound

[CLS] Token Tells Everything Needed for Training-free Efficient MLLMs cites this paper.

[CLS] Token Tells Everything Needed for Training-free Efficient MLLMs An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 10

Resolution: unresolved; no resolver link, observed 2026-08-11T20:23:18.491097Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c065c839-cc18-4729-a2f2-bd9c412471da · inbound

Enhancing Multimodal Large Language Models Complex Reason via Similarity Computation cites this paper.

Enhancing Multimodal Large Language Models Complex Reason via Similarity Computation An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 5

Resolution: unresolved; no resolver link, observed 2026-08-11T16:45:44.677283Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2096eee0-1f88-45ef-b2c8-ede51a9c376a · inbound

SoftVQ-VAE: Efficient 1-Dimensional Continuous Tokenizer cites this paper.

SoftVQ-VAE: Efficient 1-Dimensional Continuous Tokenizer An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 12

Resolution: unresolved; no resolver link, observed 2026-08-11T15:33:13.065992Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation fd6205dd-cbcc-42d1-ba3b-e588e9721631 · inbound

LLaVA Steering: Visual Instruction Tuning with 500x Fewer Parameters through Modality Linear Representation-Steering cites this paper.

LLaVA Steering: Visual Instruction Tuning with 500x Fewer Parameters through Modality Linear Representation-Steering An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 10

Resolution: unresolved; no resolver link, observed 2026-08-11T14:13:24.425557Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f29064d1-35f5-4560-81fa-fbeecb340d41 · inbound

Next Token Prediction Towards Multimodal Intelligence: A Comprehensive Survey cites this paper.

Next Token Prediction Towards Multimodal Intelligence: A Comprehensive Survey An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 56

Resolution: unresolved; no resolver link, observed 2026-08-11T14:59:01.350120Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b0a3ddb1-62de-458c-86b4-b28932ba393f · inbound

ST$^3$: Accelerating Multimodal Large Language Model by Spatial-Temporal Visual Token Trimming cites this paper.

ST$^3$: Accelerating Multimodal Large Language Model by Spatial-Temporal Visual Token Trimming An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 10

Resolution: unresolved; no resolver link, observed 2026-08-10T23:39:18.051553Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 5d19ddcd-8134-4d6a-bcfb-06d07fa22729 · inbound

Survey on Question Answering over Visually Rich Documents: Methods, Challenges, and Trends cites this paper.

Survey on Question Answering over Visually Rich Documents: Methods, Challenges, and Trends An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 15

Resolution: unresolved; no resolver link, observed 2026-08-10T22:17:27.472844Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d03d0134-b4ca-4520-9a68-69b6cf1a6e6e · inbound

What Kind of Visual Tokens Do We Need? Training-free Visual Token Pruning for Multi-modal Large Language Models from the Perspective of Graph cites this paper.

What Kind of Visual Tokens Do We Need? Training-free Visual Token Pruning for Multi-modal Large Language Models from the Perspective of Graph An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 8

Resolution: unresolved; no resolver link, observed 2026-08-10T22:20:45.651882Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 23577145-6b84-427d-b468-7100293c016a · inbound

LongViTU: Instruction Tuning for Long-Form Video Understanding cites this paper.

LongViTU: Instruction Tuning for Long-Form Video Understanding An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 9

Resolution: unresolved; no resolver link, observed 2026-08-10T21:23:57.853494Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2a5d5a1f-e4c3-41a6-96a5-2e4d0b889939 · inbound

AdaFV: Rethinking of Visual-Language alignment for VLM acceleration cites this paper.

AdaFV: Rethinking of Visual-Language alignment for VLM acceleration An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 3

Resolution: unresolved; no resolver link, observed 2026-08-10T20:00:12.568815Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ec529799-0e10-4035-b845-b9cc040f1b06 · inbound

FlexSelect: Flexible Token Selection for Efficient Long Video Understanding cites this paper.

FlexSelect: Flexible Token Selection for Efficient Long Video Understanding An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 4

Resolution: unresolved; no resolver link, observed 2026-08-07T11:59:05.780757Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6dc287d1-f569-467c-a16f-9564fbd18c74 · inbound

GreedyPrune: Retenting Critical Visual Token Set for Large Vision Language Models cites this paper.

GreedyPrune: Retenting Critical Visual Token Set for Large Vision Language Models An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 9

Resolution: unresolved; no resolver link, observed 2026-08-07T00:42:53.610866Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation eef22555-b744-4038-8e00-8a46b63f3888 · inbound

AuroraLong: Bringing RNNs Back to Efficient Open-Ended Video Understanding cites this paper.

AuroraLong: Bringing RNNs Back to Efficient Open-Ended Video Understanding An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 14

Resolution: unresolved; no resolver link, observed 2026-08-06T20:29:48.568606Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c3b71c14-67de-4ccb-8b88-9e3581a2ebf3 · inbound

ReGATE: Learning Faster and Better with Fewer Tokens in MLLMs cites this paper.

ReGATE: Learning Faster and Better with Fewer Tokens in MLLMs An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 8

Resolution: verified exact; arxiv_id, observed 2026-05-19T03:22:01.165421Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation b6e18987-f9be-456d-8b71-e8b2d7397529 · inbound

SQAP-VLA: A Synergistic Quantization-Aware Pruning Framework for High-Performance Vision-Language-Action Models cites this paper.

SQAP-VLA: A Synergistic Quantization-Aware Pruning Framework for High-Performance Vision-Language-Action Models An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 22

Resolution: unresolved; no resolver link, observed 2026-08-04T19:47:08.908760Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9f662f41-9e89-4102-bdfd-12827c0e3094 · inbound

VLN-Cache: Enabling Token Caching for VLN Models with Visual/Semantic Dynamics Awareness cites this paper.

VLN-Cache: Enabling Token Caching for VLN Models with Visual/Semantic Dynamics Awareness An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 26

Resolution: verified exact; arxiv_id, observed 2026-05-15T14:51:08.157390Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation fa624b16-8128-423d-b260-c53fb5913a90 · inbound

UIPress: Bringing Optical Token Compression to UI-to-Code Generation cites this paper.

UIPress: Bringing Optical Token Compression to UI-to-Code Generation An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 5

Resolution: verified exact; arxiv_id, observed 2026-05-11T07:00:59.145920Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation d2dc88b3-4605-468a-be51-2b4b43c1932e · inbound

Why and When Visual Token Pruning Fails? A Study on Relevant Visual Information Shift in MLLMs Decoding cites this paper.

Why and When Visual Token Pruning Fails? A Study on Relevant Visual Information Shift in MLLMs Decoding An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 9

Resolution: verified exact; arxiv_id, observed 2026-05-11T11:31:01.435400Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation 28558941-b8c5-4d6e-bbee-b078bb5e5ee3 · inbound

MergeTok: Unified Continuous and Discrete Visual Tokenization via Token Merging cites this paper.

MergeTok: Unified Continuous and Discrete Visual Tokenization via Token Merging An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 5

Resolution: verified exact; arxiv_id, observed 2026-06-28T22:42:46.050833Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation e041da95-08fc-4b5b-9c5a-86ac38731e6e · inbound

Late-Layer Fusion is Enough: Dual-Path Vision Token Routing for Multimodal Large Language Models under Visual Saturation cites this paper.

Late-Layer Fusion is Enough: Dual-Path Vision Token Routing for Multimodal Large Language Models under Visual Saturation An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 5

Resolution: metadata mismatch; arxiv_id, observed 2026-06-27T16:41:03.346578Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation 6e8e6ee6-fd67-4474-a552-bad8bf3905f7 · inbound

BVS: Bayesian Visual Search with Multimodal Large Language Model for Fine-grained Perception cites this paper.

BVS: Bayesian Visual Search with Multimodal Large Language Model for Fine-grained Perception An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 193

Resolution: unresolved; no resolver link, observed 2026-07-12T04:17:40.198357Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a8b5b3c8-cc53-4915-913d-0bbd6b662b3b · inbound

AVA-VLM: Adaptive Visual Attention-Vision Language Model for In-the-Wild Construction Site Monitoring cites this paper.

AVA-VLM: Adaptive Visual Attention-Vision Language Model for In-the-Wild Construction Site Monitoring An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 10

Resolution: verified exact; local_arxiv, observed 2026-07-08T22:35:40.573081Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation 0dbb35f5-9537-4245-861d-e21d3fa3da83 · inbound

What to Keep, What to Forget: A Rate--Distortion View of Memory Compaction in LLMs and Agents cites this paper.

What to Keep, What to Forget: A Rate--Distortion View of Memory Compaction in LLMs and Agents An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 17

Resolution: verified exact; local_arxiv, observed 2026-07-10T01:36:43.980109Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation 8c008f01-dadb-4eac-8bac-6ab703a44618 · inbound

Seeing the End at Step Zero: Accelerating Diffusion MLLMs via MLP Sparsity-Aware Truncation cites this paper.

Seeing the End at Step Zero: Accelerating Diffusion MLLMs via MLP Sparsity-Aware Truncation An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 7

Resolution: unresolved; no resolver link, observed 2026-08-02T01:48:49.788167Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 449bde39-58e9-4d03-a245-820be3235440 · inbound

LayoutLite: Token-Level Implicit Layout Analysis for Efficient Document OCR cites this paper.

LayoutLite: Token-Level Implicit Layout Analysis for Efficient Document OCR An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 2024

Resolution: unresolved; no resolver link, observed 2026-08-01T05:34:45.516112Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 90366d26-3421-42e0-8ded-37882f2138cf · inbound

ORCA: ORgan-Centroid Aggregation for Training-Free 3D CT Visual Token Compression cites this paper.

ORCA: ORgan-Centroid Aggregation for Training-Free 3D CT Visual Token Compression An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 6

Resolution: unresolved; no resolver link, observed 2026-08-04T00:46:04.099639Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 818cb237-c809-4edf-a4ac-13d5b653a7be · inbound

CRAFT: Compression via Recursive Adaptive Fusion of Video Tokens for Vision-Language Models cites this paper.

CRAFT: Compression via Recursive Adaptive Fusion of Video Tokens for Vision-Language Models An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 6

Resolution: unresolved; no resolver link, observed 2026-08-04T23:46:51.546718Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b3db6853-5d03-492e-8259-085086b20159 · inbound

SlimVLM: Sensitivity-aware Dynamic Structured Pruning with Adaptive Visual Token Selection for Efficient Vision-Language Models cites this paper.

SlimVLM: Sensitivity-aware Dynamic Structured Pruning with Adaptive Visual Token Selection for Efficient Vision-Language Models An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 11

Resolution: unresolved; no resolver link, observed 2026-08-05T16:41:28.812887Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 54c530f6-4381-45b3-8282-d55da1520d61 · inbound

Thinking With Tools, Not With Pixels: Tool Calls as Text Scaffolds for Visual Reasoning cites this paper.

Thinking With Tools, Not With Pixels: Tool Calls as Text Scaffolds for Visual Reasoning An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models

Reference 3

Resolution: unresolved; no resolver link, observed 2026-08-11T12:53:13.935715Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.