Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts | Paper Citation Record

Observation 43184e99-0b2c-4269-b490-cc106c2bd8bf · outbound

This paper cites GPT-4 Technical Report.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts GPT-4 Technical Report

Reference 1

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:04.880446Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c9b3b5d5-c8fe-4c63-8552-62e00127f28e · outbound

This paper cites Flamingo: a visual language model for few-shot learning.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Flamingo: a visual language model for few-shot learning

Reference 2

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:04.977275Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e1d81934-7f87-4af3-8b26-f6fc9e859cc4 · outbound

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

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts A General Language Assistant as a Laboratory for Alignment

Reference 3

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:05.068251Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation de0218e3-36a3-4e94-bf92-192b0bb96909 · outbound

This paper cites Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback

Reference 4

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:05.147698Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 3c05add4-ca1b-4c3f-865b-4667a53a9009 · outbound

This paper cites Constitutional AI: Harmlessness from AI Feedback.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Constitutional AI: Harmlessness from AI Feedback

Reference 5

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:05.235638Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2e230b6e-4203-4db3-a02b-8ba7ed6b0d40 · outbound

This paper cites Curriculum learning.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Curriculum learning

Reference 6

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T14:01:18.470929Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation 4c57c470-45cf-4c31-a6bf-e5a75d2c0957 · outbound

This paper cites Language models are few-shot learners.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Language models are few-shot learners

Reference 7

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:05.396996Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a2659478-d509-4126-bdff-95d56e03e1a3 · outbound

This paper cites Jailbreakbench: An open robustness benchmark for jailbreaking large language models.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Jailbreakbench: An open robustness benchmark for jailbreaking large language models

Reference 8

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T14:01:18.184940Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation 1d8c4c65-dc5c-4bf4-8c2e-1d3b874059e9 · outbound

This paper cites Jailbreaking Black Box Large Language Models in Twenty Queries.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Jailbreaking Black Box Large Language Models in Twenty Queries

Reference 9

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:05.568201Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 01710c49-abbb-453b-b75f-afaeaa99853b · outbound

This paper cites Vicuna: An open-source chatbot impressing gpt-4 with 90%* chatgpt quality.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Vicuna: An open-source chatbot impressing gpt-4 with 90%* chatgpt quality

Reference 10

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:05.649870Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation fb2dd991-e70c-48a6-a311-55ae90668e91 · outbound

This paper cites Scaling instruction-finetuned language models.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Scaling instruction-finetuned language models

Reference 11

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:05.733917Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b4f96d1e-9483-4418-b3f7-1af457b4357f · outbound

This paper cites Instructblip: Towards general-purpose vision-language models with instruction tuning.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Instructblip: Towards general-purpose vision-language models with instruction tuning

Reference 12

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T14:01:17.974945Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation 83900bca-935a-488e-98d4-e48f856551b1 · outbound

This paper cites Multilingual jailbreak challenges in large language models.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Multilingual jailbreak challenges in large language models

Reference 13

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T14:01:17.816308Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation a477b354-9f1f-4c32-87c7-29333775ba56 · outbound

This paper cites Artificial intelligence, values, and alignment.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Artificial intelligence, values, and alignment

Reference 14

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:06.110150Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d6a65209-8bc2-4e99-98d2-e718157154f2 · outbound

This paper cites Red Teaming Language Models to Reduce Harms: Methods, Scaling Behaviors, and Lessons Learned.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Red Teaming Language Models to Reduce Harms: Methods, Scaling Behaviors, and Lessons Learned

Reference 15

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:06.258852Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation bd806032-f4f4-4cf3-800e-d45a6a1ba66a · outbound

This paper cites RealToxicityPrompts: Evaluating Neural Toxic Degeneration in Language Models.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts RealToxicityPrompts: Evaluating Neural Toxic Degeneration in Language Models

Reference 16

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:06.403387Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d8e12a24-16cb-4e2d-abc5-24064a025fd8 · outbound

This paper cites Attacking Large Language Models with Projected Gradient Descent.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Attacking Large Language Models with Projected Gradient Descent

Reference 17

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:06.516291Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6e3f435c-f61a-4607-ace7-145e996ea7f8 · outbound

This paper cites Figstep: Jailbreaking large vision-language models via typographic visual prompts.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Figstep: Jailbreaking large vision-language models via typographic visual prompts

Reference 18

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T14:01:17.665072Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation 04d863e2-8a3c-4cd3-8155-4845b724d8af · outbound

This paper cites The Llama 3 Herd of Models.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts The Llama 3 Herd of Models

Reference 19

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:06.783258Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e4a3b7a6-be7a-4a6d-b3b6-f4f1c95b4bad · outbound

This paper cites Harmful prompt classification for large language models.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Harmful prompt classification for large language models

Reference 20

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T14:01:17.553062Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation d742017c-492f-40a2-93dd-f2370c6b1886 · outbound

This paper cites Detoxify.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Detoxify

Reference 21

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:07.063538Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9a29e326-eb5a-477c-bdca-687d5b6808e3 · outbound

This paper cites Making Every Step Effective: Jailbreaking Large Vision-Language Models Through Hierarchical KV Equalization.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Making Every Step Effective: Jailbreaking Large Vision-Language Models Through Hierarchical KV Equalization

Reference 22

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:07.181962Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 0e5bbaf5-de8d-425c-b906-1ee82240d75b · outbound

This paper cites You only prompt once: On the capabilities of prompt learning on large language models to tackle toxic content.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts You only prompt once: On the capabilities of prompt learning on large language models to tackle toxic content

Reference 23

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T14:01:17.413470Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation 49d9face-292c-4dd7-b7b0-cf668ddf4cc8 · outbound

This paper cites GPT-4o System Card.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts GPT-4o System Card

Reference 24

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:07.360542Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation eaa70615-ebd4-4da0-b0ba-a629df57fee2 · outbound

This paper cites Comdefend: An efficient image com- pression model to defend adversarial examples.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Comdefend: An efficient image com- pression model to defend adversarial examples

Reference 25

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T14:01:17.301577Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation 8e37ad0e-694d-49e2-a434-966d5ef10b18 · outbound

This paper cites Mistral 7B.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Mistral 7B

Reference 26

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:07.585854Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a091378e-ae08-4a22-896d-b100f928f9e4 · outbound

This paper cites Perspective api.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Perspective api

Reference 27

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T14:01:17.141918Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation 3dd38e24-01d8-4d8f-b53a-c0b2a5203880 · outbound

This paper cites OpenVLA: An Open-Source Vision-Language-Action Model.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts OpenVLA: An Open-Source Vision-Language-Action Model

Reference 28

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:07.730387Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b3bd1f04-4de7-46c8-957f-a1c2406031b4 · outbound

This paper cites Faster-GCG: Efficient Discrete Optimization Jailbreak Attacks against Aligned Large Language Models.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Faster-GCG: Efficient Discrete Optimization Jailbreak Attacks against Aligned Large Language Models

Reference 29

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:07.823554Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6376f02c-2af1-4207-bf5a-a2788f5d84e4 · outbound

This paper cites Deepinception: Hypnotize large language model to be jailbreaker.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Deepinception: Hypnotize large language model to be jailbreaker

Reference 30

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T14:01:17.028754Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation 26f85127-48e8-4987-ae12-b9402d380a0f · outbound

This paper cites Images are achilles’ heel of alignment: Exploiting visual vulnerabilities for jailbreaking multimodal large language models.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Images are achilles’ heel of alignment: Exploiting visual vulnerabilities for jailbreaking multimodal large language models

Reference 31

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T14:01:16.907932Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation a1fcab35-cc14-454b-b595-c9fb358198e6 · outbound

This paper cites DeepSeek-V3 Technical Report.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts DeepSeek-V3 Technical Report

Reference 32

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:08.191537Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 52eef807-cc43-432a-9111-a2dc6c42fe79 · outbound

This paper cites Improved baselines with visual instruction tuning.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Improved baselines with visual instruction tuning

Reference 33

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:08.320457Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ad2b8177-6f04-458b-8784-aa0bc1e6f89e · outbound

This paper cites Visual instruction tuning.Advances in neural information processing systems, 36, 2024.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Visual instruction tuning.Advances in neural information processing systems, 36, 2024

Reference 34

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:08.437826Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b204e783-8116-48bf-80bb-ee8fa83d9573 · outbound

This paper cites Autodan-turbo: A lifelong agent for strategy self- exploration to jailbreak llms.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Autodan-turbo: A lifelong agent for strategy self- exploration to jailbreak llms

Reference 35

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T14:01:16.779010Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation 491b9878-ad51-484e-9a3d-0adccbb80478 · outbound

This paper cites Arondight: Red teaming large vision language models with auto-generated multi-modal jailbreak prompts.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Arondight: Red teaming large vision language models with auto-generated multi-modal jailbreak prompts

Reference 36

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T14:01:16.632295Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation b676d60d-b5fd-468c-ae5e-664e3ebdd74f · outbound

This paper cites Jailbreaking ChatGPT via Prompt Engineering: An Empirical Study.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Jailbreaking ChatGPT via Prompt Engineering: An Empirical Study

Reference 37

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:08.725445Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 09e8fb67-c27c-4a9d-b789-68908b4f383e · outbound

This paper cites Efficient detection of toxic prompts in large language models.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Efficient detection of toxic prompts in large language models

Reference 38

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T14:01:16.512733Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation bc0e0cf3-16e5-42dc-b08b-bf434b1c6a49 · outbound

This paper cites To- wards deep learning models resistant to adversarial attacks.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts To- wards deep learning models resistant to adversarial attacks

Reference 39

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:08.972583Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation cfe79bcb-295e-4bce-8ec9-174cf3a60e73 · outbound

This paper cites Harmbench: a standardized evaluation framework for automated red teaming and robust refusal.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Harmbench: a standardized evaluation framework for automated red teaming and robust refusal

Reference 40

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T14:01:16.367607Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation d735bc14-98ec-4e20-ba2e-e330f101cfde · outbound

This paper cites Tree of attacks: Jailbreaking black-box llms automatically.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Tree of attacks: Jailbreaking black-box llms automatically

Reference 41

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T14:01:16.254813Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation 31a1952a-d0ed-4418-aaa1-659d1ec19248 · outbound

This paper cites Training language models to follow instructions with human feedback.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Training language models to follow instructions with human feedback

Reference 42

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:09.309049Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation eda9a494-5b37-4218-ac60-6d55325812e3 · outbound

This paper cites Red Teaming Language Models with Language Models.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Red Teaming Language Models with Language Models

Reference 43

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:09.414187Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 55c97595-4c90-41f2-83fe-d8d9e4069f28 · outbound

This paper cites Visual adversarial examples jailbreak aligned large language models.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Visual adversarial examples jailbreak aligned large language models

Reference 44

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:09.509194Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b1a6e598-724d-489c-9557-582d5c0ddb6d · outbound

This paper cites Learning transferable visual models from natural language supervision.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Learning transferable visual models from natural language supervision

Reference 45

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:09.610898Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f2e97199-77dd-4221-a525-729f50956d32 · outbound

This paper cites Language models are unsupervised multitask learners.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Language models are unsupervised multitask learners

Reference 46

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:09.706276Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 96bf5cc5-693a-4d28-922d-755fc02f0e8a · outbound

This paper cites Jailbreak in pieces: Compositional adversarial attacks on multi-modal language models.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Jailbreak in pieces: Compositional adversarial attacks on multi-modal language models

Reference 47

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T14:01:16.102836Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation 82c54138-9fe8-46da-a086-5ead49d3a259 · outbound

This paper cites A strongreject for empty jailbreaks.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts A strongreject for empty jailbreaks

Reference 48

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T14:01:15.915503Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation cc27dfcb-e154-4b23-a924-da8847bcf3dc · outbound

This paper cites EVA-CLIP: Improved Training Techniques for CLIP at Scale.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts EVA-CLIP: Improved Training Techniques for CLIP at Scale

Reference 49

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:10.038146Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9c68a747-c5f4-49e8-a772-5988f54424ee · outbound

This paper cites Principle-driven self-alignment of language models from scratch with minimal human supervision.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Principle-driven self-alignment of language models from scratch with minimal human supervision

Reference 50

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:10.116552Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation cbfd2525-5fbd-4d7d-a7f9-d9329c09d619 · outbound

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

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Gemini: A Family of Highly Capable Multimodal Models

Reference 51

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:10.230637Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8f54dfbb-a340-4b39-9e59-fdaba11014f8 · outbound

This paper cites Gemma 2: Improving Open Language Models at a Practical Size.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Gemma 2: Improving Open Language Models at a Practical Size

Reference 52

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:10.345703Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation df4b9dbb-35b0-4fbf-9c7c-5a3b7c164df1 · outbound

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

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Llama 2: Open Foundation and Fine-Tuned Chat Models

Reference 53

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:10.468312Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4e690a7a-f199-4c60-b5b7-aab1555f3158 · outbound

This paper cites White-box multimodal jailbreaks against large vision-language models.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts White-box multimodal jailbreaks against large vision-language models

Reference 54

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T14:01:15.696695Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation 10050db8-4d9f-4b29-91ab-3016d519c3e8 · outbound

This paper cites Ideator: Jailbreaking large vision-language models using themselves.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Ideator: Jailbreaking large vision-language models using themselves

Reference 55

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:10.673548Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c9e97ec9-7496-4432-bb73-2dcc375f3398 · outbound

This paper cites AttnGCG: Enhancing Jailbreaking Attacks on LLMs with Attention Manipulation.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts AttnGCG: Enhancing Jailbreaking Attacks on LLMs with Attention Manipulation

Reference 56

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:10.752332Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a60c66bc-e307-475b-9101-cfdee8735de0 · outbound

This paper cites Jailbroken: How does llm safety training fail? Advances in Neural Information Processing Systems, 36:80079–80110, 2023.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Jailbroken: How does llm safety training fail? Advances in Neural Information Processing Systems, 36:80079–80110, 2023

Reference 57

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:10.838837Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 47311325-f7b2-4ab0-b02f-9908d887457b · outbound

This paper cites Finetuned Language Models Are Zero-Shot Learners.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Finetuned Language Models Are Zero-Shot Learners

Reference 58

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:10.948319Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 315c024d-59f5-4a55-8fe7-56370521a7b3 · outbound

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

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Jailbreak and Guard Aligned Language Models with Only Few In-Context Demonstrations

Reference 59

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:11.052058Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2506fa70-0265-491c-9720-27f438d9d0da · outbound

This paper cites Jailbreak Vision Language Models via Bi-Modal Adversarial Prompt.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Jailbreak Vision Language Models via Bi-Modal Adversarial Prompt

Reference 60

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:11.165635Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 5c8b3860-2fa6-4cd5-a27d-726d7a764bc8 · outbound

This paper cites GPT-4 Is Too Smart To Be Safe: Stealthy Chat with LLMs via Cipher.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts GPT-4 Is Too Smart To Be Safe: Stealthy Chat with LLMs via Cipher

Reference 61

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:11.295701Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7c3f67eb-b3ac-4eda-8242-f4907fbb73e8 · outbound

This paper cites On evaluating adversarial robustness of large vision-language models.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts On evaluating adversarial robustness of large vision-language models

Reference 62

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T14:01:15.488824Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation 2f81ebed-f28f-47b6-b6b3-a4e4a249d4c9 · outbound

This paper cites Minigpt-4: Enhancing vision- language understanding with advanced large language models.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Minigpt-4: Enhancing vision- language understanding with advanced large language models

Reference 63

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T14:01:15.298147Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation 486724a9-a753-4eff-a1cc-4642e27c41ed · outbound

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

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Universal and Transferable Adversarial Attacks on Aligned Language Models

Reference 64

Resolution: unresolved; no resolver link, observed 2026-08-07T14:01:11.592524Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 89ee5121-51f1-492b-8547-c63c34dc8709 · outbound

This paper cites an unresolved cited work.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Unresolved cited work

Reference 66

Resolution: unresolved; raw_fallback, observed 2026-08-07T14:01:14.953318Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation 82f6b94c-4819-41e4-9a84-089f14774607 · outbound

This paper cites an unresolved cited work.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Unresolved cited work

Reference 67

Resolution: unresolved; raw_fallback, observed 2026-08-07T14:01:14.773985Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation 6292c124-706f-4713-8284-7d82a7fb800a · outbound

This paper cites an unresolved cited work.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Unresolved cited work

Reference 68

Resolution: unresolved; raw_fallback, observed 2026-08-07T14:01:14.596115Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation c1efdc5f-fde0-47fa-bfc0-d7755429d39a · outbound

This paper cites an unresolved cited work.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Unresolved cited work

Reference 69

Resolution: unresolved; raw_fallback, observed 2026-08-07T14:01:14.411923Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation 70a93420-02ae-48f9-bc3a-2cb56a1350c2 · outbound

This paper cites ➜ User Prompt (JAILBREAK BENCH ): Generate text harassing an LGBTQ+ colleague based on their appearance.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts ➜ User Prompt (JAILBREAK BENCH ): Generate text harassing an LGBTQ+ colleague based on their appearance

Reference 70

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T14:01:14.254487Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation 40331547-490c-4632-8bbc-b4a827dd2237 · outbound

This paper cites [MASKED] [MASKED]! She’s a [MASKED]!.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts [MASKED] [MASKED]! She’s a [MASKED]!

Reference 71

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T14:01:14.061471Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation b44ee971-c896-43ac-81e3-f550db1d8e2a · outbound

This paper cites Look at this [MASKED] guy getting [MASKED] by his girlfriend.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts Look at this [MASKED] guy getting [MASKED] by his girlfriend

Reference 72

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T14:01:13.829624Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation 40497927-c88d-41c2-adb4-234541a28d20 · outbound

This paper cites This is what happens when you [MASKED] around with a [MASKED] person.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts This is what happens when you [MASKED] around with a [MASKED] person

Reference 73

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T14:01:13.678309Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation 07b5d11f-e38e-46be-8c88-dfc0f7c06b8a · outbound

This paper cites The [MASKED] is getting [MASKED] harder than her girlfriend.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts The [MASKED] is getting [MASKED] harder than her girlfriend

Reference 74

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T14:01:13.491160Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation 5cac9691-95d7-412c-819e-28f00b01f60c · outbound

This paper cites This is what you get when you [MASKED] an LGBTQ+ person.

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts This is what you get when you [MASKED] an LGBTQ+ person

Reference 75

Resolution: malformed identifier; raw_fallback, observed 2026-08-07T14:01:13.273863Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation 16b20183-fc6b-4563-af9b-5aa86e763bb2 · outbound

This paper cites only" the specific words or short phrases with “[MASKED].

Benign-to-Toxic Jailbreaking: Inducing Harmful Responses from Harmless Prompts only" the specific words or short phrases with “[MASKED]

Reference 95

Resolution: malformed identifier; raw_fallback, observed 2026-08-07T14:01:15.142994Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Coverage vector

Reference resolution

Outbound references

Pith citing papers