EVA: Editing for Versatile Alignment against Jailbreaks | Paper Citation Record

Observation 8075fa46-4b47-41c8-875c-0d6948c08d68 · outbound

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

EVA: Editing for Versatile Alignment against Jailbreaks Training language models to follow instructions with human feedback

Reference 1

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.640896Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 254a87ed-03f7-40f7-aa2b-db563a62e682 · outbound

This paper cites Direct preference optimization: Your language model is secretly a reward model,.

EVA: Editing for Versatile Alignment against Jailbreaks Direct preference optimization: Your language model is secretly a reward model,

Reference 2

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.787946Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation fa4e60f5-5187-44b7-99fb-29b10547f463 · outbound

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

EVA: Editing for Versatile Alignment against Jailbreaks Universal and Transferable Adversarial Attacks on Aligned Language Models

Reference 3

Resolution: verified exact; local_arxiv, observed 2026-07-01T14:35:47.011132Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation dec40f2f-1823-4fe1-817f-4b74a5f7255e · outbound

This paper cites Don’t say no: Jailbreaking llm by suppressing refusal,.

EVA: Editing for Versatile Alignment against Jailbreaks Don’t say no: Jailbreaking llm by suppressing refusal,

Reference 4

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.831070Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 5629d77d-dd5f-4c80-a3b6-2641077a6864 · outbound

This paper cites Autodan: Generating stealthy jailbreak prompts on aligned large language models,.

EVA: Editing for Versatile Alignment against Jailbreaks Autodan: Generating stealthy jailbreak prompts on aligned large language models,

Reference 5

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.682727Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 6eef3067-d21f-4587-bfdf-ce9c32cb7633 · outbound

This paper cites Jailbreaking black box large language models in twenty queries,.

EVA: Editing for Versatile Alignment against Jailbreaks Jailbreaking black box large language models in twenty queries,

Reference 6

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.816253Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 17f8bfe5-340d-4536-b89b-3074cc18a57c · outbound

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

EVA: Editing for Versatile Alignment against Jailbreaks Figstep: Jailbreaking large vision-language models via typographic visual prompts,

Reference 7

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.804868Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 99fd1219-c007-4303-b66d-521a0bbd0d06 · outbound

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

EVA: Editing for Versatile Alignment against Jailbreaks Images are Achilles' Heel of Alignment: Exploiting Visual Vulnerabilities for Jailbreaking Multimodal Large Language Models

Reference 8

Resolution: metadata mismatch; arxiv_id, observed 2026-07-01T14:35:47.008475Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation f1d1cb29-cdad-44cb-8a02-78af8184eba9 · outbound

This paper cites Safety fine- tuning at (almost) no cost: A baseline for vision large language models,.

EVA: Editing for Versatile Alignment against Jailbreaks Safety fine- tuning at (almost) no cost: A baseline for vision large language models,

Reference 9

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.686821Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation edd21faa-b301-409a-b884-81d53f6597f8 · outbound

This paper cites Jailguard: A universal detection framework for prompt-based attacks on llm systems,.

EVA: Editing for Versatile Alignment against Jailbreaks Jailguard: A universal detection framework for prompt-based attacks on llm systems,

Reference 10

Resolution: verified exact; doi, observed 2026-06-30T20:45:03.018737Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 4a9b2a18-9859-4b79-961d-f35ce78acd9b · outbound

This paper cites Llavaguard: Vlm-based safeguard for vision dataset curation and safety assessment,.

EVA: Editing for Versatile Alignment against Jailbreaks Llavaguard: Vlm-based safeguard for vision dataset curation and safety assessment,

Reference 11

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.637241Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 103fb7de-00c3-4940-b76e-98131d9cc377 · outbound

This paper cites Adashield: Safe- guarding multimodal large language models from structure-based attack via adaptive shield prompting,.

EVA: Editing for Versatile Alignment against Jailbreaks Adashield: Safe- guarding multimodal large language models from structure-based attack via adaptive shield prompting,

Reference 12

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.643716Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation e1e813b4-8c9e-4de6-bc1e-2145ba601229 · outbound

This paper cites An em- pirical study of catastrophic forgetting in large language models during continual fine-tuning,.

EVA: Editing for Versatile Alignment against Jailbreaks An em- pirical study of catastrophic forgetting in large language models during continual fine-tuning,

Reference 13

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.630880Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 19f30b20-2cb3-4bdd-8d4c-bf7072edd837 · outbound

This paper cites Fine-tuning aligned language models compromises safety, even when users do not intend to!.

EVA: Editing for Versatile Alignment against Jailbreaks Fine-tuning aligned language models compromises safety, even when users do not intend to!

Reference 14

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.634289Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 9e7e1a14-2c64-4475-bc8a-acce83bad80d · outbound

This paper cites Tricking llms into disobedience: Formalizing, analyzing, and detecting jailbreaks,.

EVA: Editing for Versatile Alignment against Jailbreaks Tricking llms into disobedience: Formalizing, analyzing, and detecting jailbreaks,

Reference 15

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.782195Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 76e60fcb-b0a9-428d-951c-54904456d190 · outbound

This paper cites On the robustness of large multimodal models against image adversarial attacks,.

EVA: Editing for Versatile Alignment against Jailbreaks On the robustness of large multimodal models against image adversarial attacks,

Reference 16

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.702925Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 0d8141a6-4fdb-4267-a011-7152b50833f8 · outbound

This paper cites Refusal in language models is mediated by a single direction.

EVA: Editing for Versatile Alignment against Jailbreaks Refusal in language models is mediated by a single direction

Reference 17

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.678338Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 1d05397d-7a04-40f9-b82f-2d2f8584ce6f · outbound

This paper cites Finding safety neurons in large language models,.

EVA: Editing for Versatile Alignment against Jailbreaks Finding safety neurons in large language models,

Reference 18

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.691925Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 7dfd8acf-722f-4473-bf55-36e1bbb7534f · outbound

This paper cites Shaping the safety boundaries: Understanding and defending against jail- breaks in large language models,.

EVA: Editing for Versatile Alignment against Jailbreaks Shaping the safety boundaries: Understanding and defending against jail- breaks in large language models,

Reference 19

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.695362Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 2706bad6-f453-4657-af62-87815d8c8405 · outbound

This paper cites On information and sufficiency,.

EVA: Editing for Versatile Alignment against Jailbreaks On information and sufficiency,

Reference 20

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.699053Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 7b040890-31ac-4d6b-8ee7-7260a3328edd · outbound

This paper cites Mass-editing memory in a transformer,.

EVA: Editing for Versatile Alignment against Jailbreaks Mass-editing memory in a transformer,

Reference 21

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.706818Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation f6fd88d7-b817-45bf-9f50-ec84a88f4eee · outbound

This paper cites Easyedit: An easy-to-use knowledge editing framework for large language models,.

EVA: Editing for Versatile Alignment against Jailbreaks Easyedit: An easy-to-use knowledge editing framework for large language models,

Reference 22

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.767619Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation c0916ad6-d53f-44e2-8706-8d24eac17354 · outbound

This paper cites DELMAN: Dynamic defense against large language model jailbreaking with model editing,.

EVA: Editing for Versatile Alignment against Jailbreaks DELMAN: Dynamic defense against large language model jailbreaking with model editing,

Reference 23

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.757781Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 5a4e2b7d-814b-4c00-bbbb-3d31c2f99a02 · outbound

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

EVA: Editing for Versatile Alignment against Jailbreaks Visual adversarial examples jailbreak aligned large language models,

Reference 24

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.646640Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation f4b2cffb-a7fb-4687-9df9-bff54fd2badd · outbound

This paper cites Mm- safetybench: A benchmark for safety evaluation of multimodal large language models,.

EVA: Editing for Versatile Alignment against Jailbreaks Mm- safetybench: A benchmark for safety evaluation of multimodal large language models,

Reference 25

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.775242Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation f51be089-360e-4733-80db-9c413e7f07ea · outbound

This paper cites High-resolution image synthesis with latent diffusion models.

EVA: Editing for Versatile Alignment against Jailbreaks High-resolution image synthesis with latent diffusion models

Reference 26

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.660423Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 8e8ee226-3533-4457-8f39-719e5f36339e · outbound

This paper cites Multitrust: A comprehensive benchmark towards trustworthy multimodal large language mod- els,.

EVA: Editing for Versatile Alignment against Jailbreaks Multitrust: A comprehensive benchmark towards trustworthy multimodal large language mod- els,

Reference 27

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.845169Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 4f1d7404-2d35-4287-8052-6c18e851e589 · outbound

This paper cites Lora: Low-rank adaptation of large language models.

EVA: Editing for Versatile Alignment against Jailbreaks Lora: Low-rank adaptation of large language models

Reference 28

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.746419Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 3003a607-2ece-4a10-b278-016be2c4b16d · outbound

This paper cites Defending large language models against jailbreak attacks via layer-specific editing,.

EVA: Editing for Versatile Alignment against Jailbreaks Defending large language models against jailbreak attacks via layer-specific editing,

Reference 29

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.745005Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 709d77b2-2701-4492-b365-54afcf78fa84 · outbound

This paper cites Defending against alignment- breaking attacks via robustly aligned llm,.

EVA: Editing for Versatile Alignment against Jailbreaks Defending against alignment- breaking attacks via robustly aligned llm,

Reference 30

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.834279Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 30f07997-f0f3-416d-98c8-2c536e063e03 · outbound

This paper cites Baseline defenses for adversarial attacks against aligned language models.

EVA: Editing for Versatile Alignment against Jailbreaks Baseline defenses for adversarial attacks against aligned language models

Reference 31

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.841699Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 290da67f-d653-4185-a081-71a02cf1a11d · outbound

This paper cites Robust prompt optimization for defending language models against jailbreaking attacks,.

EVA: Editing for Versatile Alignment against Jailbreaks Robust prompt optimization for defending language models against jailbreaking attacks,

Reference 32

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.856608Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation e7b95ca8-9bec-4532-93af-5972668973b6 · outbound

This paper cites Safedecoding: Defending against jailbreak attacks via safety- aware decoding,.

EVA: Editing for Versatile Alignment against Jailbreaks Safedecoding: Defending against jailbreak attacks via safety- aware decoding,

Reference 33

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.820007Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 9227303d-daf8-4f2a-b503-f9c213903487 · outbound

This paper cites Improving alignment and robustness with circuit breakers,.

EVA: Editing for Versatile Alignment against Jailbreaks Improving alignment and robustness with circuit breakers,

Reference 34

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.822905Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 1482cd86-b1fe-4b2e-ad82-40bdd6ae556f · outbound

This paper cites Cross-modality information check for detecting jailbreaking in multimodal large language models,.

EVA: Editing for Versatile Alignment against Jailbreaks Cross-modality information check for detecting jailbreaking in multimodal large language models,

Reference 35

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.827276Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 7419ef02-27f1-4ee7-a6ac-fec04b2ea847 · outbound

This paper cites Fast model editing at scale.

EVA: Editing for Versatile Alignment against Jailbreaks Fast model editing at scale

Reference 36

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.798203Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 6ddcd638-bddc-45a4-b29d-d081d97b56a0 · outbound

This paper cites Editing factual knowledge in language models.

EVA: Editing for Versatile Alignment against Jailbreaks Editing factual knowledge in language models

Reference 37

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.812887Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation d00c895b-4b36-4201-9264-efc607c04c49 · outbound

This paper cites Modifying memories in transformer models,.

EVA: Editing for Versatile Alignment against Jailbreaks Modifying memories in transformer models,

Reference 38

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.801071Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 83cece88-f10d-45e8-8e9a-1ec6e95234d9 · outbound

This paper cites Plug-and-play adaptation for continuously-updated qa,.

EVA: Editing for Versatile Alignment against Jailbreaks Plug-and-play adaptation for continuously-updated qa,

Reference 39

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.794903Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 0c06983a-b1f6-4675-bd14-2759d1d88b0a · outbound

This paper cites an unresolved cited work.

EVA: Editing for Versatile Alignment against Jailbreaks Unresolved cited work

Reference 40

Resolution: unresolved; raw_fallback, observed 2026-07-07T19:24:06.784872Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation a4b5a223-c7a4-48a4-a34e-eabd4d763ce5 · outbound

This paper cites Locating and editing factual associations in gpt,.

EVA: Editing for Versatile Alignment against Jailbreaks Locating and editing factual associations in gpt,

Reference 41

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.791685Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 0771ef56-5f38-49f9-9fbe-5016314a7c95 · outbound

This paper cites Detoxifying large language models via knowledge editing,.

EVA: Editing for Versatile Alignment against Jailbreaks Detoxifying large language models via knowledge editing,

Reference 42

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.808956Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 196b1798-6d5e-4e81-b407-4adc1a65076c · outbound

This paper cites Safeint: Shielding large language models from jailbreak attacks via safety-aware represen- tation intervention,.

EVA: Editing for Versatile Alignment against Jailbreaks Safeint: Shielding large language models from jailbreak attacks via safety-aware represen- tation intervention,

Reference 43

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.862963Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 419cddc2-f0f5-4d7b-8f6b-f173f022c366 · outbound

This paper cites SafeLLM: Unlearning Harmful Outputs from Large Language Models against Jailbreak Attacks.

EVA: Editing for Versatile Alignment against Jailbreaks SafeLLM: Unlearning Harmful Outputs from Large Language Models against Jailbreak Attacks

Reference 44

Resolution: verified exact; arxiv_id, observed 2026-07-01T14:35:47.030620Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 2ed8ea68-3d80-4015-a443-11f4067454ac · outbound

This paper cites Safety alignment via constrained knowledge unlearn- ing,.

EVA: Editing for Versatile Alignment against Jailbreaks Safety alignment via constrained knowledge unlearn- ing,

Reference 45

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.741970Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation e9e6ca0d-073c-4b79-96ff-ceffef485bb9 · outbound

This paper cites Towards automated circuit discovery for mechanistic interpretability,.

EVA: Editing for Versatile Alignment against Jailbreaks Towards automated circuit discovery for mechanistic interpretability,

Reference 46

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.852744Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation aa616b5d-aa62-4220-8488-59ae4b646260 · outbound

This paper cites Attention is all you need.

EVA: Editing for Versatile Alignment against Jailbreaks Attention is all you need

Reference 47

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.663326Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 11bd81ee-d262-45a1-9674-2e584995b034 · outbound

This paper cites Transformer feed- forward layers are key-value memories.

EVA: Editing for Versatile Alignment against Jailbreaks Transformer feed- forward layers are key-value memories

Reference 48

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.671385Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 32044bf0-ad4a-47c6-b03c-74c3c340f815 · outbound

This paper cites Transformer feed-forward layers build predictions by promoting concepts in the vocabulary space,.

EVA: Editing for Versatile Alignment against Jailbreaks Transformer feed-forward layers build predictions by promoting concepts in the vocabulary space,

Reference 49

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.726194Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 5cb1fb02-624f-406f-a5e4-1ab83ae6459e · outbound

This paper cites GLU Variants Improve Transformer.

EVA: Editing for Versatile Alignment against Jailbreaks GLU Variants Improve Transformer

Reference 50

Resolution: verified exact; local_arxiv, observed 2026-07-01T14:35:47.021364Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 63f75ce1-bfef-4f39-a595-ea3891bc6e08 · outbound

This paper cites GPT-4o System Card.

EVA: Editing for Versatile Alignment against Jailbreaks GPT-4o System Card

Reference 51

Resolution: verified exact; local_arxiv, observed 2026-07-01T14:35:47.035526Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 895cbbc3-f6f4-4c05-9a1c-12f6a0dad55b · outbound

This paper cites Sparsevlm: Vi- sual token sparsification for efficient vision-language model infer- ence,.

EVA: Editing for Versatile Alignment against Jailbreaks Sparsevlm: Vi- sual token sparsification for efficient vision-language model infer- ence,

Reference 52

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.747985Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation fcec43ba-b740-49b5-8a94-6139b3c9b131 · outbound

This paper cites Generic attention-model ex- plainability for interpreting bi-modal and encoder-decoder trans- formers,.

EVA: Editing for Versatile Alignment against Jailbreaks Generic attention-model ex- plainability for interpreting bi-modal and encoder-decoder trans- formers,

Reference 53

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.770927Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 385ea8e4-afda-49da-8e75-e09700436aef · outbound

This paper cites Evaluating object hallucination in large vision-language models,.

EVA: Editing for Versatile Alignment against Jailbreaks Evaluating object hallucination in large vision-language models,

Reference 54

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.729527Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation bb6c03ea-b366-43fc-9832-a6ba11456c4a · outbound

This paper cites Rewriting a deep generative model,.

EVA: Editing for Versatile Alignment against Jailbreaks Rewriting a deep generative model,

Reference 55

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.733278Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 4047abc1-ac5f-4cbd-9e1b-7420eef718ef · outbound

This paper cites Foundation.

EVA: Editing for Versatile Alignment against Jailbreaks Foundation

Reference 56

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.866647Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation b2215723-e17c-45af-8927-6e050aae8831 · outbound

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

EVA: Editing for Versatile Alignment against Jailbreaks Harmbench: a standardized evaluation framework for automated red teaming and robust refusal,

Reference 57

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.778991Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 37ec42cd-9115-4f21-a071-06550e0c4301 · outbound

This paper cites Jailbreakbench: An open robustness benchmark for jail- breaking large language models,.

EVA: Editing for Versatile Alignment against Jailbreaks Jailbreakbench: An open robustness benchmark for jail- breaking large language models,

Reference 58

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.624568Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation e8499363-47b1-497f-adf5-1eab1e2cf320 · outbound

This paper cites Catastrophic jailbreak of open-source llms via exploiting generation,.

EVA: Editing for Versatile Alignment against Jailbreaks Catastrophic jailbreak of open-source llms via exploiting generation,

Reference 59

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.711240Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 273d566e-2c2e-4f2b-a924-5bb8814ef01e · outbound

This paper cites Judging llm-as-a-judge with mt- bench and chatbot arena,.

EVA: Editing for Versatile Alignment against Jailbreaks Judging llm-as-a-judge with mt- bench and chatbot arena,

Reference 60

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.627545Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation dec13a08-7d66-4146-bbea-150be9d1cbfb · outbound

This paper cites Boolq: Exploring the surprising difficulty of natural yes/no questions,.

EVA: Editing for Versatile Alignment against Jailbreaks Boolq: Exploring the surprising difficulty of natural yes/no questions,

Reference 61

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.621759Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 1c33657e-8201-4628-a882-86c04db7b34f · outbound

This paper cites Squad: 100,000+ questions for machine comprehension of text,.

EVA: Editing for Versatile Alignment against Jailbreaks Squad: 100,000+ questions for machine comprehension of text,

Reference 62

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.714986Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 9d878e2f-ac7d-4342-84c1-2883ae01fafb · outbound

This paper cites Mutual: A dataset for multi-turn dialogue reasoning,.

EVA: Editing for Versatile Alignment against Jailbreaks Mutual: A dataset for multi-turn dialogue reasoning,

Reference 63

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.718485Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation de7f7d33-cccd-4ea1-be47-bd1828510d93 · outbound

This paper cites Introduction to the conll-2003 shared task: Language-independent named entity recognition,.

EVA: Editing for Versatile Alignment against Jailbreaks Introduction to the conll-2003 shared task: Language-independent named entity recognition,

Reference 64

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.722535Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 3c369a44-9950-47e8-a432-37353c7a58a6 · outbound

This paper cites The pascal recognising textual entailment challenge,.

EVA: Editing for Versatile Alignment against Jailbreaks The pascal recognising textual entailment challenge,

Reference 65

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.761155Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation c09724b4-aefb-4bf8-bdbf-19f3f80338be · outbound

This paper cites Training Verifiers to Solve Math Word Problems.

EVA: Editing for Versatile Alignment against Jailbreaks Training Verifiers to Solve Math Word Problems

Reference 66

Resolution: verified exact; local_arxiv, observed 2026-07-01T14:35:47.013642Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation d1360cbd-eb3d-43bc-8f78-c21d9ec45008 · outbound

This paper cites Recursive deep models for semantic compo- sitionality over a sentiment treebank,.

EVA: Editing for Versatile Alignment against Jailbreaks Recursive deep models for semantic compo- sitionality over a sentiment treebank,

Reference 67

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.736286Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation fcfe467c-ddc1-4b65-bcf6-65d6acda0905 · outbound

This paper cites Samsum corpus: A human-annotated dialogue dataset for abstractive summariza- tion,.

EVA: Editing for Versatile Alignment against Jailbreaks Samsum corpus: A human-annotated dialogue dataset for abstractive summariza- tion,

Reference 68

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.754308Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 6ac330c9-6d9a-49aa-91c5-d07f31371753 · outbound

This paper cites Rouge: A package for automatic evaluation of summaries,.

EVA: Editing for Versatile Alignment against Jailbreaks Rouge: A package for automatic evaluation of summaries,

Reference 69

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.743692Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation d4e0e9ad-0470-45b4-96e0-4e8ee369c14f · outbound

This paper cites Mm-vet v2: A challenging benchmark to evaluate large multimodal models for integrated capabilities,.

EVA: Editing for Versatile Alignment against Jailbreaks Mm-vet v2: A challenging benchmark to evaluate large multimodal models for integrated capabilities,

Reference 70

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.656547Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 4a2f3f0d-3bd1-437e-8998-111a4c7d2936 · outbound

This paper cites Mmmu: A massive multi-discipline multimodal understanding and reasoning benchmark for expert agi,.

EVA: Editing for Versatile Alignment against Jailbreaks Mmmu: A massive multi-discipline multimodal understanding and reasoning benchmark for expert agi,

Reference 71

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.720469Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 187132fa-c66e-4765-9c4a-3548e78a154c · outbound

This paper cites Are we on the right way for evaluating large vision-language models?.

EVA: Editing for Versatile Alignment against Jailbreaks Are we on the right way for evaluating large vision-language models?

Reference 72

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.764327Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation bf974ba2-c3a5-4e81-8179-4b0d6f27ca36 · outbound

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

EVA: Editing for Versatile Alignment against Jailbreaks Llama 2: Open Foundation and Fine-Tuned Chat Models

Reference 73

Resolution: verified exact; local_arxiv, observed 2026-07-01T14:35:47.019023Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation c0b44040-18f3-4332-9857-a6d4cfc6dda8 · outbound

This paper cites Mistral 7B.

EVA: Editing for Versatile Alignment against Jailbreaks Mistral 7B

Reference 74

Resolution: verified exact; local_arxiv, observed 2026-07-01T14:35:47.027376Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation ed49cd05-d763-447e-b125-71cf8d7b44e8 · outbound

This paper cites The Llama 3 Herd of Models.

EVA: Editing for Versatile Alignment against Jailbreaks The Llama 3 Herd of Models

Reference 75

Resolution: verified exact; local_arxiv, observed 2026-07-01T14:35:47.033081Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation ba87f120-10e2-4338-88bd-11fba1e91c66 · outbound

This paper cites Qwen2.5 Technical Report.

EVA: Editing for Versatile Alignment against Jailbreaks Qwen2.5 Technical Report

Reference 76

Resolution: verified exact; local_arxiv, observed 2026-07-01T14:35:47.024632Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation a9933b21-80fb-47ac-90dc-118bb4b1ccaa · outbound

This paper cites Visual instruction tuning.

EVA: Editing for Versatile Alignment against Jailbreaks Visual instruction tuning

Reference 77

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.649344Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation f0b6e546-5fc9-429e-a741-a605fb1e191e · outbound

This paper cites Qwen2.5-VL Technical Report.

EVA: Editing for Versatile Alignment against Jailbreaks Qwen2.5-VL Technical Report

Reference 78

Resolution: verified exact; local_arxiv, observed 2026-07-01T14:35:47.021667Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 93a5e94a-5570-4b47-b599-4be69a726d35 · outbound

This paper cites InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiency.

EVA: Editing for Versatile Alignment against Jailbreaks InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiency

Reference 79

Resolution: verified exact; local_arxiv, observed 2026-07-01T14:35:47.038302Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 78c08467-ad18-4b8a-bce6-1b4a506192c7 · outbound

This paper cites an unresolved cited work.

EVA: Editing for Versatile Alignment against Jailbreaks Unresolved cited work

Reference 80

Resolution: unresolved; raw_fallback, observed 2026-07-07T19:24:06.738816Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 5677bb5d-2284-4e5b-ad02-74e6894e9220 · outbound

This paper cites A vector space model for automatic indexing.

EVA: Editing for Versatile Alignment against Jailbreaks A vector space model for automatic indexing

Reference 81

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.848933Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 9251fb0a-eb57-41d5-9dab-3d94c9cee571 · outbound

This paper cites On adaptive attacks to adversarial example defenses,.

EVA: Editing for Versatile Alignment against Jailbreaks On adaptive attacks to adversarial example defenses,

Reference 82

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.750988Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation bd578b27-f52d-49b2-a0b3-065eb97d94df · outbound

This paper cites Principal component analy- sis,.

EVA: Editing for Versatile Alignment against Jailbreaks Principal component analy- sis,

Reference 83

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.859888Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 729d498d-ec39-47b6-8991-2e9952377dc9 · outbound

This paper cites Divergence measures based on the shannon entropy,.

EVA: Editing for Versatile Alignment against Jailbreaks Divergence measures based on the shannon entropy,

Reference 84

Resolution: verified fuzzy; raw_fallback, observed 2026-07-07T19:24:06.837175Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Coverage vector

Reference resolution

Outbound references

Pith citing papers