Paper Citation Record · LEDGER

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models

As of 18 August 2026, this Paper Citation Record lists 82 of 82 outbound references and 2 inbound Pith citation observations for arXiv:2504.18053.

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

pith.paper-citation-record.v1
2504.18053 v2

Coverage vector

measured 82 of 82 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-08-16T10:29:59.372879Z

measured 84 of 84 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 2 of 2 inbound itemization

Pith citing papers itemized under the disclosed page cap.

Source: paper_references, paper_reference_links, observed 2026-08-05T20:18:54.813869Z

measured 0 of 1 external citation measurements

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

Source: pith, observed 2026-08-05T20:19:00.210071Z

Reference resolution

82 of 82 outbound references displayed

verified exact4
verified fuzzy1
unresolved77
parse uncertain0
malformed identifier0
metadata mismatch0

External citation measurements

No source-named external measurement is stored.

Outbound references

Observation e10a668f-ff63-47f7-ba33-17cc35019d98 · outbound

This paper cites Gemini: A Family of Highly Capable Multimodal Models.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Gemini: A Family of Highly Capable Multimodal Models

Reference 1

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:58.971332Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 920bde59-3b8c-4ef0-abcb-d00a8846cda6 · outbound

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

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models A General Language Assistant as a Laboratory for Alignment

Reference 2

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:58.977026Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d2265866-541a-48bd-85de-94793a13743a · outbound

This paper cites MT-Bench-101: A Fine-Grained Benchmark for Evaluating Large Language Models in Multi-Turn Dialogues.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models MT-Bench-101: A Fine-Grained Benchmark for Evaluating Large Language Models in Multi-Turn Dialogues

Reference 3

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:58.982687Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f4c62138-73d6-4df0-a052-cdfc3f598138 · outbound

This paper cites Qwen Technical Report.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Qwen Technical Report

Reference 4

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:58.987874Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6453a9a0-6f92-4783-83e2-6679b4e1673b · outbound

This paper cites Qwen-VL: A Versatile Vision-Language Model for Understanding, Localization, Text Reading, and Beyond.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Qwen-VL: A Versatile Vision-Language Model for Understanding, Localization, Text Reading, and Beyond

Reference 5

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:58.993072Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 89008e3b-3ba3-444e-840c-ba318a3372ba · outbound

This paper cites an unresolved cited work.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Unresolved cited work

Reference 6

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:58.999131Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 81888494-e5c1-45b1-ac99-8f792c28af5c · outbound

This paper cites an unresolved cited work.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Unresolved cited work

Reference 7

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.004473Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7b3f03d4-9325-4e14-a3e5-866a670df179 · outbound

This paper cites an unresolved cited work.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Unresolved cited work

Reference 8

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.009251Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b03ca546-c6e5-487e-8df9-f9d557dfe0ea · outbound

This paper cites DRESS: Instructing Large Vision-Language Models to Align and Interact with Humans via Natural Language Feedback.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models DRESS: Instructing Large Vision-Language Models to Align and Interact with Humans via Natural Language Feedback

Reference 9

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.013789Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2fa4920d-fa2b-4f04-913a-3930d14a598b · outbound

This paper cites How Far Are We to GPT-4V? Closing the Gap to Commercial Multimodal Models with Open-Source Suites.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models How Far Are We to GPT-4V? Closing the Gap to Commercial Multimodal Models with Open-Source Suites

Reference 10

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.018850Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 321581cb-67b2-47ef-8cd4-a10e27f89f2d · outbound

This paper cites The Llama 3 Herd of Models.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models The Llama 3 Herd of Models

Reference 11

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.023665Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e642e2ec-2e7f-45d9-abb8-d06edb607040 · outbound

This paper cites Beyond Bounding Box: Multimodal Knowledge Learning for Object Detection.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Beyond Bounding Box: Multimodal Knowledge Learning for Object Detection

Reference 12

Resolution: verified exact; local_arxiv, observed 2026-08-16T10:30:00.329619Z

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 ce2a40d0-0c34-462c-a33a-fc5b4b9bac9c · outbound

This paper cites MME: A Comprehensive Evaluation Benchmark for Multimodal Large Language Models.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models MME: A Comprehensive Evaluation Benchmark for Multimodal Large Language Models

Reference 13

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.033134Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 96e609c9-bff4-4bd4-8966-09c7f91e03bd · outbound

This paper cites an unresolved cited work.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Unresolved cited work

Reference 14

Resolution: unresolved; raw_fallback, observed 2026-08-16T10:30:00.903628Z

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 5ecdc404-efe4-4b23-ba54-eefb57098d32 · outbound

This paper cites FigStep: Jailbreaking Large Vision-Language Models via Typographic Visual Prompts.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models FigStep: Jailbreaking Large Vision-Language Models via Typographic Visual Prompts

Reference 15

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.043297Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d03a804c-7366-4312-a850-c99569fefb0a · outbound

This paper cites Eyes Closed, Safety On: Protecting Multimodal LLMs via Image-to-Text Transformation.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Eyes Closed, Safety On: Protecting Multimodal LLMs via Image-to-Text Transformation

Reference 16

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.048143Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 5fb28a5f-78a2-4e0f-a880-986184cf5263 · outbound

This paper cites LlavaGuard: An Open VLM-based Framework for Safeguarding Vision Datasets and Models.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models LlavaGuard: An Open VLM-based Framework for Safeguarding Vision Datasets and Models

Reference 17

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.053086Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d9ec4943-39cd-4658-9888-fe8e4564d69f · outbound

This paper cites Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations

Reference 18

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.058063Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ac5a4299-35e7-47b3-a200-ecc7701696ad · outbound

This paper cites an unresolved cited work.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Unresolved cited work

Reference 19

Resolution: unresolved; raw_fallback, observed 2026-08-16T10:30:00.887901Z

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 7b4ded2a-73be-4701-a3db-19fe7e9c64e3 · outbound

This paper cites an unresolved cited work.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Unresolved cited work

Reference 20

Resolution: unresolved; raw_fallback, observed 2026-08-16T10:30:00.871326Z

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 18e6de6f-c778-4753-be36-d24bb3a5ca3f · outbound

This paper cites an unresolved cited work.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Unresolved cited work

Reference 21

Resolution: unresolved; raw_fallback, observed 2026-08-16T10:30:00.855235Z

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 34b1071f-a8e4-4e70-8624-c4bb8f6a79eb · outbound

This paper cites Red Teaming Visual Language Models.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Red Teaming Visual Language Models

Reference 22

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.078230Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7201b62d-4f52-4eaf-bfce-ccce2a59e927 · outbound

This paper cites GraphReader: Building Graph-based Agent to Enhance Long-Context Abilities of Large Language Models.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models GraphReader: Building Graph-based Agent to Enhance Long-Context Abilities of Large Language Models

Reference 23

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.083147Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 754297b5-7d17-4251-9528-590a75c576a1 · outbound

This paper cites 2D-DPO: Scaling Direct Preference Optimization with 2-Dimensional Supervision.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models 2D-DPO: Scaling Direct Preference Optimization with 2-Dimensional Supervision

Reference 24

Resolution: verified exact; local_arxiv, observed 2026-08-16T10:30:00.183863Z

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 1ada7bc9-fb86-482b-8448-842b150b5413 · outbound

This paper cites an unresolved cited work.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Unresolved cited work

Reference 25

Resolution: unresolved; raw_fallback, observed 2026-08-16T10:30:00.839602Z

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 7d65da0c-0b9f-4c82-a02c-41a53ee8ecfe · outbound

This paper cites MOSSBench: Is Your Multimodal Language Model Oversensitive to Safe Queries?.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models MOSSBench: Is Your Multimodal Language Model Oversensitive to Safe Queries?

Reference 26

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.096856Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b5c0eedf-30c6-4bb5-8717-2426ff33417c · outbound

This paper cites Images are Achilles' Heel of Alignment: Exploiting Visual Vulnerabilities for Jailbreaking Multimodal Large Language Models.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Images are Achilles' Heel of Alignment: Exploiting Visual Vulnerabilities for Jailbreaking Multimodal Large Language Models

Reference 27

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.102631Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e598a9fa-a54b-44f7-9d58-ff8a97d5c83b · outbound

This paper cites an unresolved cited work.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Unresolved cited work

Reference 28

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.107689Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d9023dbc-02ea-4a9c-89b1-50ee4d1943ef · outbound

This paper cites Improved Baselines with Visual Instruction Tuning.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Improved Baselines with Visual Instruction Tuning

Reference 29

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.112050Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9bcdd601-1b87-48b2-96d2-d6190127beb1 · outbound

This paper cites Iterative Length-Regularized Direct Preference Optimization: A Case Study on Improving 7B Language Models to GPT-4 Level.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Iterative Length-Regularized Direct Preference Optimization: A Case Study on Improving 7B Language Models to GPT-4 Level

Reference 30

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.116751Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a83073d4-886f-4d45-9300-16090a46f2ce · outbound

This paper cites MM-SafetyBench: A Benchmark for Safety Evaluation of Multimodal Large Language Models.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models MM-SafetyBench: A Benchmark for Safety Evaluation of Multimodal Large Language Models

Reference 31

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.121834Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1ffa7a5d-a5a6-49bf-b707-3849dbe35b9c · outbound

This paper cites MMBench: Is Your Multi-modal Model an All-around Player?.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models MMBench: Is Your Multi-modal Model an All-around Player?

Reference 32

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.127382Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 11d60464-b52b-4b51-8f82-038bce3c60ec · outbound

This paper cites Safety Alignment for Vision Language Models.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Safety Alignment for Vision Language Models

Reference 33

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.132998Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7ff11a08-0e70-4305-b211-caff5fc8c12c · outbound

This paper cites An Image Is Worth 1000 Lies: Adversarial Transferability across Prompts on Vision-Language Models.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models An Image Is Worth 1000 Lies: Adversarial Transferability across Prompts on Vision-Language Models

Reference 34

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.138779Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2af09d0a-8eca-4f13-a722-e47bb60bcdb5 · outbound

This paper cites JailBreakV: A Benchmark for Assessing the Robustness of MultiModal Large Language Models against Jailbreak Attacks.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models JailBreakV: A Benchmark for Assessing the Robustness of MultiModal Large Language Models against Jailbreak Attacks

Reference 35

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.144283Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e56c47f5-4c82-4002-9b79-4be3af792083 · outbound

This paper cites Task-Agnostic Detector for Insertion-Based Backdoor Attacks.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Task-Agnostic Detector for Insertion-Based Backdoor Attacks

Reference 36

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.150098Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4502462c-5347-4648-a415-9bacf4086e6a · outbound

This paper cites TrojVLM: Backdoor Attack Against Vision Language Models.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models TrojVLM: Backdoor Attack Against Vision Language Models

Reference 37

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.155145Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8d47211b-1b88-4235-b13a-ce695deeeced · outbound

This paper cites an unresolved cited work.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Unresolved cited work

Reference 38

Resolution: unresolved; raw_fallback, observed 2026-08-16T10:30:00.810769Z

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 57336205-d557-4c31-ae11-e3d48cba736e · outbound

This paper cites an unresolved cited work.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Unresolved cited work

Reference 39

Resolution: unresolved; raw_fallback, observed 2026-08-16T10:30:00.795555Z

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 e294cb0b-df38-476e-a139-1425823e6f1b · outbound

This paper cites an unresolved cited work.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Unresolved cited work

Reference 40

Resolution: unresolved; raw_fallback, observed 2026-08-16T10:30:00.780044Z

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 0804183a-8e3d-4dce-975c-b169cabd75af · outbound

This paper cites HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal

Reference 41

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.173229Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 12bb00a4-1813-4fc4-8a45-c886f92bc5dd · outbound

This paper cites an unresolved cited work.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Unresolved cited work

Reference 42

Resolution: unresolved; raw_fallback, observed 2026-08-16T10:30:00.764280Z

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 dd5f31ad-3bb6-4dc2-94e5-5b7d306f9f80 · outbound

This paper cites an unresolved cited work.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Unresolved cited work

Reference 43

Resolution: unresolved; raw_fallback, observed 2026-08-16T10:30:00.745782Z

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 21770cc5-fb1b-4a52-94aa-67cf869777d5 · outbound

This paper cites an unresolved cited work.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Unresolved cited work

Reference 44

Resolution: unresolved; raw_fallback, observed 2026-08-16T10:30:00.729901Z

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 1f79dd50-80cf-4678-b343-0fd7ac6a8045 · outbound

This paper cites an unresolved cited work.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Unresolved cited work

Reference 45

Resolution: unresolved; raw_fallback, observed 2026-08-16T10:30:00.714621Z

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 4fcc994a-0ebf-4d38-b009-53ee54075103 · outbound

This paper cites an unresolved cited work.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Unresolved cited work

Reference 46

Resolution: unresolved; raw_fallback, observed 2026-08-16T10:30:00.699931Z

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 2f5b8a44-d811-4e7f-8b5e-1b1eb58037e7 · outbound

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

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Kosmos-2: Grounding Multimodal Large Language Models to the World

Reference 47

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.201777Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9184d978-542a-4a2f-b138-26d8d6d3740c · outbound

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

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models MLLM-Protector: Ensuring MLLM's Safety without Hurting Performance

Reference 48

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.206118Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation bbb4aa37-b4ca-468e-be4b-fdc5f2d6b759 · outbound

This paper cites an unresolved cited work.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Unresolved cited work

Reference 49

Resolution: unresolved; raw_fallback, observed 2026-08-16T10:30:00.684205Z

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 ec910c25-2acc-46c2-800f-63bcc9600935 · outbound

This paper cites Vision-LLMs Can Fool Themselves with Self-Generated Typographic Attacks.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Vision-LLMs Can Fool Themselves with Self-Generated Typographic Attacks

Reference 50

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.215890Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1543a04e-4dfd-4404-96d1-3604b461394c · outbound

This paper cites an unresolved cited work.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Unresolved cited work

Reference 51

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.221062Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8676ace5-dbef-4beb-b254-dc046f8c6d6c · outbound

This paper cites Manning, Stefano Ermon, and Chelsea Finn.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Manning, Stefano Ermon, and Chelsea Finn

Reference 52

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.225273Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation eaf57920-81b8-40d6-8a07-4ee4aa92fd2f · outbound

This paper cites Privacy-Aware Visual Language Models.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Privacy-Aware Visual Language Models

Reference 53

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.230542Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation eb05220e-52e1-426e-9039-a42c89752871 · outbound

This paper cites Proximal Policy Optimization Algorithms.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Proximal Policy Optimization Algorithms

Reference 54

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.235695Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation da963c02-f5ce-4f58-875d-b4827d04ef15 · outbound

This paper cites an unresolved cited work.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Unresolved cited work

Reference 55

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.240877Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation aedcd443-720b-453f-876c-57183f47f159 · outbound

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

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Llama 2: Open Foundation and Fine-Tuned Chat Models

Reference 56

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.245927Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 828726e6-328c-4a8a-9ecf-29d4a5569377 · outbound

This paper cites How Many Unicorns Are in This Image? A Safety Evaluation Benchmark for Vision LLMs.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models How Many Unicorns Are in This Image? A Safety Evaluation Benchmark for Vision LLMs

Reference 57

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.250602Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b7136024-4a0f-43f2-aba8-a883317281a3 · outbound

This paper cites Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution

Reference 58

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.255476Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation cc2f6355-a0dd-41f2-b6f7-3a0ce5fc7f91 · outbound

This paper cites InferAligner: Inference-Time Alignment for Harmlessness through Cross-Model Guidance.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models InferAligner: Inference-Time Alignment for Harmlessness through Cross-Model Guidance

Reference 59

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.260099Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 843de802-8fef-408b-a208-006bb50228d5 · outbound

This paper cites Safe Inputs but Unsafe Output: Benchmarking Cross-modality Safety Alignment of Large Vision-Language Model.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Safe Inputs but Unsafe Output: Benchmarking Cross-modality Safety Alignment of Large Vision-Language Model

Reference 60

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.265087Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 09b4c3f0-9275-4e1e-be3e-070b3dbb123a · outbound

This paper cites AdaShield: Safeguarding Multimodal Large Language Models from Structure-based Attack via Adaptive Shield Prompting.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models AdaShield: Safeguarding Multimodal Large Language Models from Structure-based Attack via Adaptive Shield Prompting

Reference 61

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.269688Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a667b95c-f065-40ed-9633-ac9b522fb94b · outbound

This paper cites ConceptMath: A Bilingual Concept-wise Benchmark for Measuring Mathematical Reasoning of Large Language Models.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models ConceptMath: A Bilingual Concept-wise Benchmark for Measuring Mathematical Reasoning of Large Language Models

Reference 62

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.274420Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8d74e4e2-32eb-40f5-a239-97160921d891 · outbound

This paper cites an unresolved cited work.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Unresolved cited work

Reference 63

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.279014Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b2e5acda-cf48-41a7-9b76-1b5b6242cba4 · outbound

This paper cites an unresolved cited work.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Unresolved cited work

Reference 64

Resolution: unresolved; raw_fallback, observed 2026-08-16T10:30:00.630807Z

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 d145773c-6b39-4f39-abeb-280470ae7451 · outbound

This paper cites MiniCPM-V: A GPT-4V Level MLLM on Your Phone.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models MiniCPM-V: A GPT-4V Level MLLM on Your Phone

Reference 65

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.288629Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 56615a39-b363-45a7-bc3b-f271aa4115f5 · outbound

This paper cites an unresolved cited work.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Unresolved cited work

Reference 66

Resolution: unresolved; raw_fallback, observed 2026-08-16T10:30:00.615290Z

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 eae13ea7-2212-4157-840a-b421c95babeb · outbound

This paper cites an unresolved cited work.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Unresolved cited work

Reference 67

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.299257Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2da26ba0-46fb-42b6-87c2-d4b4ccee9f55 · outbound

This paper cites an unresolved cited work.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Unresolved cited work

Reference 68

Resolution: unresolved; raw_fallback, observed 2026-08-16T10:30:00.590160Z

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 a7eefa9e-03c1-4ac7-8d79-0db22f695ebf · outbound

This paper cites an unresolved cited work.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Unresolved cited work

Reference 69

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.307889Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2f6ad466-de22-4fd4-b232-bb27ac31ab8e · outbound

This paper cites an unresolved cited work.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Unresolved cited work

Reference 70

Resolution: verified exact; doi, observed 2026-08-16T10:29:59.413268Z

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 ff745b42-e919-413e-917f-7e9ea2559edd · outbound

This paper cites RATT: A Thought Structure for Coherent and Correct LLM Reasoning.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models RATT: A Thought Structure for Coherent and Correct LLM Reasoning

Reference 71

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.316516Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 5d881f95-076f-404d-9802-470245054a55 · outbound

This paper cites JailGuard: A Universal Detection Framework for LLM Prompt-based Attacks.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models JailGuard: A Universal Detection Framework for LLM Prompt-based Attacks

Reference 72

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.321345Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e13a54c2-2ee4-45ef-bc19-39809676faea · outbound

This paper cites an unresolved cited work.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Unresolved cited work

Reference 73

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.326126Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 54153688-e3aa-42d4-8156-c713dfaa01b6 · outbound

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

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models SPA-VL: A Comprehensive Safety Preference Alignment Dataset for Vision Language Model

Reference 74

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.331297Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 05d84306-305f-433d-905b-8d7060698a2c · outbound

This paper cites The First to Know: How Token Distributions Reveal Hidden Knowledge in Large Vision-Language Models?.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models The First to Know: How Token Distributions Reveal Hidden Knowledge in Large Vision-Language Models?

Reference 75

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.336577Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1b1d460c-baa6-413c-81ce-23e3d40836db · outbound

This paper cites Xing, Hao Zhang, Joseph E.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Xing, Hao Zhang, Joseph E

Reference 76

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.343407Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 89248376-634e-4e7a-8192-0a7a736028b8 · outbound

This paper cites On the existence of a trojaned twin model.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models On the existence of a trojaned twin model

Reference 77

Resolution: verified fuzzy; raw_fallback, observed 2026-08-16T10:30:00.564068Z

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 bfdbd885-c1c4-424c-9c28-2c4bb19c5599 · outbound

This paper cites Regression and Forecasting of U.S. Stock Returns Based on LSTM.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Regression and Forecasting of U.S. Stock Returns Based on LSTM

Reference 78

Resolution: verified exact; local_arxiv, observed 2026-08-16T10:29:59.471023Z

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 2fe48823-906f-4092-b583-8ed8049551ba · outbound

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

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Safety Fine-Tuning at (Almost) No Cost: A Baseline for Vision Large Language Models

Reference 79

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.358383Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2758b88b-1a74-458d-af51-b1f601befb28 · outbound

This paper cites Universal and Transferable Adversarial Attacks on Aligned Language Models.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models Universal and Transferable Adversarial Attacks on Aligned Language Models

Reference 80

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.363129Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ccdc77a8-50d3-49fb-9e8f-9ad8a0fc093d · outbound

This paper cites online" 'onlinestring :=.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models online" 'onlinestring :=

Reference 81

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.367786Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 5aa8a94e-b269-4913-bb55-f2b4b998dc63 · outbound

This paper cites write newline.

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models write newline

Reference 82

Resolution: unresolved; no resolver link, observed 2026-08-16T10:29:59.372879Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Pith citing papers

Observation 701e583d-73bb-48cf-91f2-fe35ab599e05 · inbound

MM-BrowseComp: A Comprehensive Benchmark for Multimodal Browsing Agents cites this paper.

MM-BrowseComp: A Comprehensive Benchmark for Multimodal Browsing Agents DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models

Reference 22

Resolution: verified exact; local_arxiv, observed 2026-08-05T20:19:00.281226Z

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 e41a74eb-bb88-48a3-a0e8-309109c9ae6d · inbound

Attention Misses Visual Risk: Risk-Adaptive Steering for Multimodal Safety Alignment cites this paper.

Attention Misses Visual Risk: Risk-Adaptive Steering for Multimodal Safety Alignment DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models

Reference 10

Resolution: unresolved; no resolver link, observed 2026-08-04T09:47:23.933962Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.