Paper Citation Record · LEDGER

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap

As of 19 August 2026, this Paper Citation Record lists 30 of 30 outbound references and 1 inbound Pith citation observation for arXiv:2505.24208.

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

pith.paper-citation-record.v1
2505.24208 v1

Coverage vector

measured 30 of 30 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-08-07T12:37:22.900435Z

measured 31 of 31 standing notices

One-hop event checks from named stored sources.

Source: scholarly_work_events, retraction_status_cache, observed 2026-08-18T06:34:40.430872+00:00

measured 1 of 1 inbound itemization

Pith citing papers itemized under the disclosed page cap.

Source: paper_references, paper_reference_links, observed 2026-05-10T18:03:36.805784Z

measured 0 of 1 external citation measurements

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

Source: arxiv_reference, observed 2026-05-11T05:35:59.023089Z

Reference resolution

30 of 30 outbound references displayed

verified exact0
verified fuzzy1
unresolved29
parse uncertain0
malformed identifier0
metadata mismatch0

External citation measurements

No source-named external measurement is stored.

Outbound references

Observation 8ab9883d-a676-4add-b462-e240a7892f62 · outbound

This paper cites GPT-4 Technical Report.

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap GPT-4 Technical Report

Reference 1

Resolution: unresolved; no resolver link, observed 2026-08-07T12:37:20.087864Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f14bc96e-8981-434c-bc43-e9e382469cef · outbound

This paper cites ✓ Guide me through the process of strangle a bird until it stops moving.

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap ✓ Guide me through the process of strangle a bird until it stops moving

Reference 3

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T12:37:23.384490Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation 9607eea5-77ce-4ffb-a95d-3db29813f1e8 · outbound

This paper cites The revolution of multimodal large language models: a survey.arXiv preprint arXiv:2402.12451,.

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap The revolution of multimodal large language models: a survey.arXiv preprint arXiv:2402.12451,

Reference 4

Resolution: unresolved; no resolver link, observed 2026-08-07T12:37:20.321312Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f06c6d6c-d71e-4590-b3e3-9b141f91c912 · outbound

This paper cites MiniGPT-v2: large language model as a unified interface for vision-language multi-task learning.

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap MiniGPT-v2: large language model as a unified interface for vision-language multi-task learning

Reference 5

Resolution: unresolved; no resolver link, observed 2026-08-07T12:37:20.446909Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d6506708-7542-411a-90f0-fac087b3d32c · outbound

This paper cites CoCA: Regaining Safety-awareness of Multimodal Large Language Models with Constitutional Calibration.

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap CoCA: Regaining Safety-awareness of Multimodal Large Language Models with Constitutional Calibration

Reference 8

Resolution: unresolved; no resolver link, observed 2026-08-07T12:37:20.727581Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 76781e91-6e51-4482-a6bc-524960b990f1 · outbound

This paper cites The Llama 3 Herd of Models.

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap The Llama 3 Herd of Models

Reference 9

Resolution: unresolved; no resolver link, observed 2026-08-07T12:37:20.806425Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 53e7085f-7024-4d79-b1b9-55ac3cd7e269 · outbound

This paper cites HallusionBench: An Advanced Diagnostic Suite for Entangled Language Hallucination and Visual Illusion in Large Vision-Language Models.

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap HallusionBench: An Advanced Diagnostic Suite for Entangled Language Hallucination and Visual Illusion in Large Vision-Language Models

Reference 10

Resolution: unresolved; no resolver link, observed 2026-08-07T12:37:20.873566Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c11fcb7d-a043-4238-9988-5fe90ad14551 · outbound

This paper cites Deciphering Cross-Modal Alignment in Large Vision-Language Models with Modality Integration Rate.

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap Deciphering Cross-Modal Alignment in Large Vision-Language Models with Modality Integration Rate

Reference 11

Resolution: unresolved; no resolver link, observed 2026-08-07T12:37:20.984510Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 05916946-9f08-4808-8aed-d87770cbb58d · outbound

This paper cites Certifying LLM Safety against Adversarial Prompting.

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap Certifying LLM Safety against Adversarial Prompting

Reference 13

Resolution: unresolved; no resolver link, observed 2026-08-07T12:37:21.159164Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1364f870-658b-4d4c-9fad-112e776f57a4 · outbound

This paper cites SEED-Bench: Benchmarking Multimodal LLMs with Generative Comprehension.

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap SEED-Bench: Benchmarking Multimodal LLMs with Generative Comprehension

Reference 14

Resolution: unresolved; no resolver link, observed 2026-08-07T12:37:21.246083Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b52951d0-c59e-45a9-943d-c1c4cccf67d6 · outbound

This paper cites Unraveling and Mitigating Safety Alignment Degradation of Vision-Language Models.

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap Unraveling and Mitigating Safety Alignment Degradation of Vision-Language Models

Reference 15

Resolution: unresolved; no resolver link, observed 2026-08-07T12:37:21.360286Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 72ca9c3c-08d6-4d15-babe-287dc7417621 · outbound

This paper cites Learn to Explain: Multimodal Reasoning via Thought Chains for Science Question Answering.

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap Learn to Explain: Multimodal Reasoning via Thought Chains for Science Question Answering

Reference 16

Resolution: unresolved; no resolver link, observed 2026-08-07T12:37:21.439899Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 21c92529-f865-4992-afa6-e748b1f35487 · outbound

This paper cites ChartQA: A Benchmark for Question Answering about Charts with Visual and Logical Reasoning.

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap ChartQA: A Benchmark for Question Answering about Charts with Visual and Logical Reasoning

Reference 17

Resolution: unresolved; no resolver link, observed 2026-08-07T12:37:21.559574Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d7c8e246-fc69-4cf1-ab10-0bcfebdc19ad · outbound

This paper cites Kosmos-2: Grounding Multimodal Large Language Models to the World.

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap Kosmos-2: Grounding Multimodal Large Language Models to the World

Reference 18

Resolution: unresolved; no resolver link, observed 2026-08-07T12:37:21.691913Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 373bac8e-af7d-4313-97ae-a9f57bf5bb85 · outbound

This paper cites MLLM-Protector: Ensuring MLLM's Safety without Hurting Performance.

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap MLLM-Protector: Ensuring MLLM's Safety without Hurting Performance

Reference 19

Resolution: unresolved; no resolver link, observed 2026-08-07T12:37:21.767959Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a4393d1e-3329-4552-94e9-39b145ddd14c · outbound

This paper cites Visual Adversarial Examples Jailbreak Aligned Large Language Models.

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap Visual Adversarial Examples Jailbreak Aligned Large Language Models

Reference 20

Resolution: unresolved; no resolver link, observed 2026-08-07T12:37:21.846602Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7002ee0a-0cf9-4f58-8b30-b0490a8dc042 · outbound

This paper cites Towards VQA Models That Can Read.

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap Towards VQA Models That Can Read

Reference 21

Resolution: unresolved; no resolver link, observed 2026-08-07T12:37:21.942532Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9a1bcc6f-82f3-4541-834a-604b1821ab20 · outbound

This paper cites Llama 2: Open Foundation and Fine-Tuned Chat Models.

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap Llama 2: Open Foundation and Fine-Tuned Chat Models

Reference 22

Resolution: unresolved; no resolver link, observed 2026-08-07T12:37:22.035080Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b85afa86-7506-40d3-a94d-a5f5d8a0e6fd · outbound

This paper cites RLHFPoison: Reward Poisoning Attack for Reinforcement Learning with Human Feedback in Large Language Models.

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap RLHFPoison: Reward Poisoning Attack for Reinforcement Learning with Human Feedback in Large Language Models

Reference 23

Resolution: unresolved; no resolver link, observed 2026-08-07T12:37:22.138901Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 5faeabe1-ec16-46fd-8fbe-ec8a76ffd52e · outbound

This paper cites Jailbreak and Guard Aligned Language Models with Only Few In-Context Demonstrations.

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap Jailbreak and Guard Aligned Language Models with Only Few In-Context Demonstrations

Reference 24

Resolution: unresolved; no resolver link, observed 2026-08-07T12:37:22.236454Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c2634783-01b7-413c-b107-6d60b806f3f1 · outbound

This paper cites mPLUG-Owl: Modularization Empowers Large Language Models with Multimodality.

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap mPLUG-Owl: Modularization Empowers Large Language Models with Multimodality

Reference 25

Resolution: unresolved; no resolver link, observed 2026-08-07T12:37:22.338677Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1db626b9-2698-4367-afcd-c3ca4ce1ee8d · outbound

This paper cites SPA-VL: A Comprehensive Safety Preference Alignment Dataset for Vision Language Model.

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap SPA-VL: A Comprehensive Safety Preference Alignment Dataset for Vision Language Model

Reference 26

Resolution: unresolved; no resolver link, observed 2026-08-07T12:37:22.450280Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a09b9c7c-a99c-4223-8237-a7818a11bf2e · outbound

This paper cites BlueSuffix: Reinforced Blue Teaming for Vision-Language Models Against Jailbreak Attacks.

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap BlueSuffix: Reinforced Blue Teaming for Vision-Language Models Against Jailbreak Attacks

Reference 27

Resolution: unresolved; no resolver link, observed 2026-08-07T12:37:22.581493Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1334fb5a-992c-45bf-8850-eda677fa1b27 · outbound

This paper cites MiniGPT-4: Enhancing Vision-Language Understanding with Advanced Large Language Models.

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap MiniGPT-4: Enhancing Vision-Language Understanding with Advanced Large Language Models

Reference 28

Resolution: unresolved; no resolver link, observed 2026-08-07T12:37:22.691154Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation eb3ea9cb-e2c6-4341-a225-4e1d40090c0e · outbound

This paper cites Safety Fine-Tuning at (Almost) No Cost: A Baseline for Vision Large Language Models.

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap Safety Fine-Tuning at (Almost) No Cost: A Baseline for Vision Large Language Models

Reference 29

Resolution: unresolved; no resolver link, observed 2026-08-07T12:37:22.785969Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 88939420-e15b-40ee-b0d6-33a93fa31dc8 · outbound

This paper cites Microsoft COCO Captions: Data Collection and Evaluation Server.

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap Microsoft COCO Captions: Data Collection and Evaluation Server

Reference 2015

Resolution: unresolved; no resolver link, observed 2026-08-07T12:37:20.542530Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8746388f-f6cd-4b81-9bd2-6da365a75987 · outbound

This paper cites GQA: A New Dataset for Real-World Visual Reasoning and Compositional Question Answering.

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap GQA: A New Dataset for Real-World Visual Reasoning and Compositional Question Answering

Reference 2019

Resolution: unresolved; no resolver link, observed 2026-08-07T12:37:21.085852Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2efd0ce3-1b5e-4d72-af33-f45de2da836e · outbound

This paper cites A General Language Assistant as a Laboratory for Alignment.

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap A General Language Assistant as a Laboratory for Alignment

Reference 2022

Resolution: unresolved; no resolver link, observed 2026-08-07T12:37:20.145735Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 10490e61-fe9e-4378-8a94-1dcea2e4c117 · outbound

This paper cites ETA: Evaluating Then Aligning Safety of Vision Language Models at Inference Time.

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap ETA: Evaluating Then Aligning Safety of Vision Language Models at Inference Time

Reference 2023

Resolution: unresolved; no resolver link, observed 2026-08-07T12:37:20.634917Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6c8db265-1e66-4b5d-b441-c8ea788277f3 · outbound

This paper cites Constitutional AI: Harmlessness from AI Feedback.

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap Constitutional AI: Harmlessness from AI Feedback

Reference 2024

Resolution: unresolved; no resolver link, observed 2026-08-07T12:37:20.236651Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Pith citing papers

Observation 32032ed4-50be-4a95-9eab-abe186f93d66 · inbound

Mosaic: Multimodal Jailbreak against Closed-Source VLMs via Multi-View Ensemble Optimization cites this paper.

Mosaic: Multimodal Jailbreak against Closed-Source VLMs via Multi-View Ensemble Optimization Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap

Reference 39

Resolution: verified exact; arxiv_id, observed 2026-05-11T05:35:59.026216Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.