Pith. sign in

REVIEW 23 cited by

SPA-VL: A Comprehensive Safety Preference Alignment Dataset for Vision Language Model

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

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 2406.12030 v4 pith:62PRQDHY submitted 2024-06-17 cs.CV cs.AIcs.CL

Yongting Zhang , Lu Chen , Guodong Zheng , Yifeng Gao , Rui Zheng , Jinlan Fu , Zhenfei Yin , Senjie Jin
show 5 more authors
Yu Qiao Xuanjing Huang Feng Zhao Tao Gui Jing Shao
This is my paper · ORCID
classification cs.CVcs.AIcs.CL
keywords alignmentspa-vlvlmsdatasetmodelssafetylanguagepreference
verification ladder T0 review T1 audit T2 compute T3 formal

Signed reviews

No signed human review yet.

0 comments
read the original abstract

The emergence of Vision Language Models (VLMs) has brought unprecedented advances in understanding multimodal information. The combination of textual and visual semantics in VLMs is highly complex and diverse, making the safety alignment of these models challenging. Furthermore, due to the limited study on the safety alignment of VLMs, there is a lack of large-scale, high-quality datasets. To address these limitations, we propose a Safety Preference Alignment dataset for Vision Language Models named SPA-VL. In terms of breadth, SPA-VL covers 6 harmfulness domains, 13 categories, and 53 subcategories, and contains 100,788 samples of the quadruple (question, image, chosen response, rejected response). In terms of depth, the responses are collected from 12 open-source (e.g., QwenVL) and closed-source (e.g., Gemini) VLMs to ensure diversity. The construction of preference data is fully automated, and the experimental results indicate that models trained with alignment techniques on the SPA-VL dataset exhibit substantial improvements in harmlessness and helpfulness while maintaining core capabilities. SPA-VL, as a large-scale, high-quality, and diverse dataset, represents a significant milestone in ensuring that VLMs achieve both harmlessness and helpfulness.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 23 Pith papers

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

  1. SafeCap: Improving LVLM Safety with Image Captioning Reinforcement Learning

    cs.CV 2026-08 conditional novelty 7.0 of 10

    Training LVLMs to produce safety-relevant image captions before answering, with a frozen-LLM caption reward, raises multimodal safety average by up to 19 points without lowering vision utility.

  2. MDIT-Bench: Evaluating the Dual-Implicit Toxicity in Large Multimodal Models

    cs.CL 2025-05 conditional novelty 7.0 of 10

    The authors release a large multimodal benchmark showing that current LMMs struggle to detect toxicity that emerges only from combining image and text, and that many-shot toxic demonstrations further reduce their accuracy.

  3. Multimodal Reward Hacking in Reinforcement Learning

    cs.AI 2026-07 conditional novelty 6.5 of 10

    Imperfect multimodal RL rewards systematically create new failures (NRFR > RHR); scaling and answer-aware rewards help but do not eliminate hacking, and unreliable visual verifiers actively increase it.

  4. When Are Reasoning-Based Guardrails Not Efficient? ResponseGuard: A Fast Vision-Language Guard for Real-Time Moderation

    cs.CV 2026-07 conditional novelty 6.0 of 10

    A label-only 2B guard outperforms a 3B reasoning-based vision-language guard on response-harmfulness detection at about 150x lower latency, with the remaining image gap attributed to frozen encoders.

  5. GrAInS: Gradient-based Attribution for Inference-Time Steering of LLMs and VLMs

    cs.CL 2025-07 conditional novelty 6.0 of 10

    GrAInS uses Integrated Gradients to identify the most influential tokens, then builds layer-wise steering vectors that improve truthfulness, reduce hallucination, and preserve general capabilities in LLMs and VLMs.

  6. Bridging the Gap in Vision Language Models in Identifying Unsafe Concepts Across Modalities

    cs.CR 2025-07 conditional novelty 6.0 of 10

    Vision-language models consistently recognize unsafe content better from text than from images, and a simplified reinforcement learning fine-tune narrows that gap.

  7. The Man Behind the Sound: Demystifying Audio Private Attribute Profiling via Multimodal Large Language Model Agents

    cs.CR 2025-07 unverdicted novelty 6.0 of 10

    A multi-agent audio-language model framework can automatically profile private attributes, such as age, health, and income, directly from general audio recordings.

  8. GThinker: Towards General Multimodal Reasoning via Cue-Guided Rethinking

    cs.CV 2025-06 conditional novelty 6.0 of 10

    A cue-tagging and rethinking training recipe lifts a 7B multimodal model to 81.5% on M3CoT, though the gain may reflect training on the same benchmark.

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

    cs.AI 2025-05 conditional novelty 6.0 of 10

    Reducing the modality gap between image and text embeddings during LVLM pretraining reduces unsafe response rates by up to 16.3% across models and benchmarks.

  10. USB: A Comprehensive and Unified Safety Evaluation Benchmark for Multimodal Large Language Models

    cs.CR 2025-05 conditional novelty 6.0 of 10

    USB-SafeBench is a unified MLLM safety benchmark with 61 risk categories, 4 modality combinations, and dual-language vulnerability and oversensitivity tests.

  11. ShieldVLM: Safeguarding the Multimodal Implicit Toxicity via Deliberative Reasoning with LVLMs

    cs.MM 2025-05 conditional novelty 6.0 of 10

    ShieldVLM detects multimodal implicit toxicity through deliberate cross-modal reasoning, outperforming existing moderation APIs and models on the new MMIT benchmark.

  12. MORALISE: A Structured Benchmark for Moral Alignment in Visual Language Models

    cs.CV 2025-05 conditional novelty 6.0 of 10

    MORALISE is a new benchmark of 2,481 expert-annotated real image-text pairs spanning 13 moral topics, and 19 vision-language models score far worse on identifying the violated norm than on judging whether a violation ...

  13. VLSBench: Unveiling Visual Leakage in Multimodal Safety

    cs.CR 2024-11 conditional novelty 6.0 of 10

    The paper shows existing multimodal safety benchmarks leak harmful image content into text queries (VSIL), and introduces VLSBench, a 2.2k-pair leakless benchmark on which textual alignment fails and multimodal alignm...

  14. A Survey on Evaluating Quality and Trustworthiness in LLM-Generated Data

    cs.AI 2026-01 conditional novelty 5.0 of 10

    A metric-oriented survey that classifies intrinsic quality and trustworthiness metrics for LLM-generated data across six modalities and documents systematic evaluation gaps in the current literature.

  15. Seeing the Threat: Vulnerabilities in Vision-Language Models to Adversarial Attack

    cs.CL 2025-05 conditional novelty 5.0 of 10

    A two-stage evaluation framework and token-projection analysis show that LVLMs encode harmful semantic cues from images even without OCR, while remaining vulnerable to cross-modal attacks.

  16. VSCBench: Bridging the Gap in Vision-Language Model Safety Calibration

    cs.IR 2025-05 conditional novelty 5.0 of 10

    A new benchmark, VSCBench, measures oversafety and undersafety in vision-language models and shows that most models, including proprietary ones, are miscalibrated on at least one safety dimension.

  17. GuardReasoner-VL: Safeguarding VLMs via Reinforced Reasoning

    cs.AI 2025-05 conditional novelty 5.0 of 10

    GuardReasoner-VL, a 3B/7B VLM guard model trained with reasoning SFT and online RL, reports large F1 gains over existing VLM guard models on 14 safety benchmarks.

  18. Mapping User Trust in Vision Language Models: Research Landscape, Challenges, and Prospects

    cs.CV 2025-05 accept novelty 5.0 of 10

    A structured review and pilot workshop that organizes VLM trust research into a new taxonomy and finds a shortage of direct user studies.

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

    cs.CL 2025-04 conditional novelty 5.0 of 10

    Multimodal risk disentanglement, where the model breaks down threats from images and text separately, improves MLLM safety at inference and during fine-tuning.

  20. Internal Activation Revision: Safeguarding Vision Language Models Without Parameter Update

    cs.LG 2025-01 conditional novelty 5.0 of 10

    Adding contrastive activation vectors at selected layers or heads reduces jailbreak success in LLaVA and Qwen-VL models by roughly 23 to 57 percentage points across four safety benchmarks, with a modest drop in helpfu...

  21. Critic-V: VLM Critics Help Catch VLM Errors in Multimodal Reasoning

    cs.CV 2024-11 conditional novelty 5.0 of 10

    A DPO-trained VLM critic that critiques and iteratively refines a reasoning VLM improves accuracy on several multimodal benchmarks, with large gains on MathVista and RealWorldQA.

  22. A Survey of Safety on Large Vision-Language Models: Attacks, Defenses and Evaluations

    cs.CR 2025-02 conditional novelty 4.0 of 10

    A survey of LVLM safety that adds a lifecycle taxonomy and new benchmark results showing Janus-Pro-7B has weaker safety than several open-source LVLMs.

  23. Aligning Large Vision-Language Models by Deep Reinforcement Learning and Direct Preference Optimization

    cs.LG 2025-09 unverdicted

    A survey categorizing deep reinforcement learning and direct preference optimization methods for aligning large vision-language models, with tables of studies and datasets and no new experimental result.

Pith tools