Paper Citation Record · LEDGER

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection

As of 4 August 2026, this Paper Citation Record lists 37 of 37 outbound references and 0 inbound Pith citation observations for arXiv:2604.12548.

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

pith.paper-citation-record.v1
2604.12548 v1

Coverage vector

measured 37 of 37 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-05-10T15:54:24.013408Z

measured 37 of 37 standing notices

One-hop event checks from named stored sources.

Source: scholarly_work_events, retraction_status_cache, observed 2026-08-03T06:30:56.289259+00:00

measured 0 of 0 inbound itemization

Pith citing papers itemized under the disclosed page cap.

Source: paper_references, paper_reference_links

measured 0 of 1 external citation measurements

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

Source: cited_works

Reference resolution

37 of 37 outbound references displayed

verified exact8
verified fuzzy24
unresolved0
parse uncertain0
malformed identifier0
metadata mismatch5

External citation measurements

No source-named external measurement is stored.

Outbound references

Observation ff2e81a2-bc07-4ce1-994f-10bffca44805 · outbound

This paper cites Model-as-a-service (MaaS): A survey.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection Model-as-a-service (MaaS): A survey

Reference 1

Resolution: verified fuzzy; raw_fallback, observed 2026-05-17T18:02:02.746286Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation 755a6bc8-f523-400c-8dbe-df07891bf848 · outbound

This paper cites Multimodal large language models: A survey.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection Multimodal large language models: A survey

Reference 2

Resolution: verified fuzzy; raw_fallback, observed 2026-05-17T18:02:02.736069Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation 7ef0c7a5-9477-4d32-afe5-c0d26fd3cf92 · outbound

This paper cites Mixture of experts (MoE): A big data perspective.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection Mixture of experts (MoE): A big data perspective

Reference 3

Resolution: verified fuzzy; raw_fallback, observed 2026-05-17T18:02:02.789126Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation fccb413b-75e1-4a2b-a9c9-a8e50611f09a · outbound

This paper cites Safety in Large Reasoning Models: A Survey.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection Safety in Large Reasoning Models: A Survey

Reference 4

Resolution: verified exact; arxiv_id, observed 2026-05-11T09:41:00.379274Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation e2cb6d56-33ae-4da5-b50d-6b63c27a7b7b · outbound

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

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection Jailbreaking black box large language models in twenty queries

Reference 5

Resolution: verified fuzzy; raw_fallback, observed 2026-05-17T18:02:02.776819Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation 50d2e92d-6450-48e5-8377-1ca602d369a4 · outbound

This paper cites Not what you’ve signed up for: Compromising real-world LLM- integrated applications with indirect prompt injection.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection Not what you’ve signed up for: Compromising real-world LLM- integrated applications with indirect prompt injection

Reference 6

Resolution: verified fuzzy; raw_fallback, observed 2026-05-17T18:02:02.779970Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation 7311c844-7b4e-4ce5-bbe1-9ca087b68089 · outbound

This paper cites AdvPrompter: Fast Adaptive Adversarial Prompting for LLMs.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection AdvPrompter: Fast Adaptive Adversarial Prompting for LLMs

Reference 7

Resolution: metadata mismatch; arxiv_id, observed 2026-05-11T09:41:00.474344Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation 9f11f76f-e81e-4496-905d-025b4969ae8b · outbound

This paper cites Certified defenses for data poisoning attacks.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection Certified defenses for data poisoning attacks

Reference 8

Resolution: verified fuzzy; raw_fallback, observed 2026-05-17T18:02:02.783168Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation 66950a75-5670-4354-9dfa-278da53d908c · outbound

This paper cites Defending large language models against jailbreaking attacks through goal prioriti- zation.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection Defending large language models against jailbreaking attacks through goal prioriti- zation

Reference 9

Resolution: verified fuzzy; raw_fallback, observed 2026-05-17T18:02:02.782085Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation a0d34ecd-8fa9-43a5-86b9-963b12bac8f1 · outbound

This paper cites Many-shot jailbreaking.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection Many-shot jailbreaking

Reference 10

Resolution: verified fuzzy; raw_fallback, observed 2026-05-17T18:02:02.785927Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation 60765ed4-2fd4-4543-82d4-c42df9607390 · outbound

This paper cites Poisoning language models during instruction tuning.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection Poisoning language models during instruction tuning

Reference 11

Resolution: verified fuzzy; raw_fallback, observed 2026-05-17T18:02:02.763969Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation 75988f09-8143-4832-9e70-7b30bcbb1180 · outbound

This paper cites Universal adversarial triggers for attacking and analyzing nlp.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection Universal adversarial triggers for attacking and analyzing nlp

Reference 12

Resolution: verified fuzzy; raw_fallback, observed 2026-05-17T18:02:02.798072Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation 07bb2040-79c4-4d8b-8084-1686e0b0b7d5 · outbound

This paper cites Auto- Prompt: Eliciting knowledge from language models with automatically generated prompts.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection Auto- Prompt: Eliciting knowledge from language models with automatically generated prompts

Reference 13

Resolution: verified fuzzy; raw_fallback, observed 2026-05-17T18:02:02.804722Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation e228b997-e501-4bda-a1f6-c32d01b0fc79 · outbound

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

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection Universal and Transferable Adversarial Attacks on Aligned Language Models

Reference 14

Resolution: verified exact; local_arxiv, observed 2026-05-11T09:41:00.437700Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation 9ecfd9f7-59e3-4629-8560-2ab84e0268bf · outbound

This paper cites Promptrobust: Towards evaluating the robustness of large language models on adversarial prompts.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection Promptrobust: Towards evaluating the robustness of large language models on adversarial prompts

Reference 15

Resolution: verified fuzzy; raw_fallback, observed 2026-05-17T18:02:02.801237Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation 2ce9acb9-1790-42da-9d54-5affe3a6c8af · outbound

This paper cites Red teaming visual language models.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection Red teaming visual language models

Reference 16

Resolution: verified fuzzy; raw_fallback, observed 2026-05-17T18:02:02.784734Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation da67828f-94dc-4725-9469-2999c1b10128 · outbound

This paper cites arXiv preprint arXiv:2503.11519 (2025).

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection arXiv preprint arXiv:2503.11519 (2025)

Reference 17

Resolution: metadata mismatch; arxiv_id, observed 2026-05-11T09:41:00.456076Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation 6f4fdde2-2c2e-47b0-8e86-f356c9abc9e5 · outbound

This paper cites The Attacker Moves Second: Stronger Adaptive Attacks Bypass Defenses Against Llm Jailbreaks and Prompt Injections.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection The Attacker Moves Second: Stronger Adaptive Attacks Bypass Defenses Against Llm Jailbreaks and Prompt Injections

Reference 18

Resolution: verified exact; arxiv_id, observed 2026-05-16T18:53:24.731592Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation 2fefa364-0629-4ad5-8811-f9d5aa02e6cf · outbound

This paper cites ShieldLearner: A New Paradigm for Jailbreak Attack Defense in LLMs.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection ShieldLearner: A New Paradigm for Jailbreak Attack Defense in LLMs

Reference 19

Resolution: verified exact; arxiv_id, observed 2026-05-11T09:41:00.350285Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation e607288b-a2b4-448b-beb5-10c6b5ac97f3 · outbound

This paper cites The hidden risks of large reasoning models: A safety assessment of R1.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection The hidden risks of large reasoning models: A safety assessment of R1

Reference 20

Resolution: verified fuzzy; raw_fallback, observed 2026-05-17T18:02:02.761539Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation 7f1cea53-f874-4c94-9901-1654983e25e3 · outbound

This paper cites Sugar- coated poison: Benign generation unlocks jailbreaking.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection Sugar- coated poison: Benign generation unlocks jailbreaking

Reference 21

Resolution: verified fuzzy; raw_fallback, observed 2026-05-17T18:02:02.758396Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation 40e707ae-1686-42f8-a653-8d729ee20f00 · outbound

This paper cites Token-Efficient Prompt Injection Attack: Provoking Cessation in LLM Reasoning via Adaptive Token Compression.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection Token-Efficient Prompt Injection Attack: Provoking Cessation in LLM Reasoning via Adaptive Token Compression

Reference 22

Resolution: verified exact; arxiv_id, observed 2026-05-11T09:41:00.466332Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation 4e3d6dd6-5798-4141-9e5c-f3a71aa59f40 · outbound

This paper cites Poisoning attacks on llms require a near-constant number of poison samples.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection Poisoning attacks on llms require a near-constant number of poison samples

Reference 23

Resolution: verified exact; arxiv_id, observed 2026-05-11T09:41:00.366632Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation 45a44b9c-ac91-4639-b9b0-4b9f1d851381 · outbound

This paper cites Adversarial training for free!.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection Adversarial training for free!

Reference 24

Resolution: verified fuzzy; raw_fallback, observed 2026-05-17T18:02:02.795329Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation c1d2003a-a1d7-4ef3-bcfe-9b2bb49aa241 · outbound

This paper cites Amnesiac machine learning.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection Amnesiac machine learning

Reference 25

Resolution: verified fuzzy; raw_fallback, observed 2026-05-17T18:02:02.751809Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation bb794607-65bb-4515-ab46-ec0f478c43a5 · outbound

This paper cites Membership inference attacks against machine learning models.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection Membership inference attacks against machine learning models

Reference 26

Resolution: verified fuzzy; raw_fallback, observed 2026-05-17T18:02:02.792570Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation 3a0cd059-ed16-48ae-84d2-45ea7e2dab99 · outbound

This paper cites Ackerman and Nina Panickssery.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection Ackerman and Nina Panickssery

Reference 27

Resolution: metadata mismatch; arxiv_id, observed 2026-05-11T09:41:00.406514Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation 3e2646fb-9dee-4921-9c2b-956ecad09058 · outbound

This paper cites SelfDefend: LLMs can defend themselves against jailbreaking in a practical manner.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection SelfDefend: LLMs can defend themselves against jailbreaking in a practical manner

Reference 28

Resolution: verified fuzzy; raw_fallback, observed 2026-05-17T18:02:02.807796Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation 05402d42-fc70-4b25-af19-5c5434c6e40c · outbound

This paper cites A survey on in-context learning.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection A survey on in-context learning

Reference 29

Resolution: verified fuzzy; raw_fallback, observed 2026-05-17T18:02:02.765062Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation fc29c32f-22c0-42d5-9478-04311f7110ef · outbound

This paper cites Red teaming language models with language models.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection Red teaming language models with language models

Reference 30

Resolution: verified fuzzy; raw_fallback, observed 2026-05-17T18:02:02.761842Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation 10013782-d979-4cdf-a8d6-d3d5b5a3d67c · outbound

This paper cites Strip: A defence against trojan attacks on deep neural networks.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection Strip: A defence against trojan attacks on deep neural networks

Reference 31

Resolution: verified fuzzy; raw_fallback, observed 2026-05-17T18:02:02.740652Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation bf5b1655-afab-417b-88e9-f3c814500768 · outbound

This paper cites A Stitch in Time Saves Nine: Detecting and Mitigating Hallucinations of LLMs by Validating Low-Confidence Generation.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection A Stitch in Time Saves Nine: Detecting and Mitigating Hallucinations of LLMs by Validating Low-Confidence Generation

Reference 32

Resolution: metadata mismatch; arxiv_id, observed 2026-05-11T09:41:00.398146Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation b5fd5c6b-01a9-4196-9edf-fe856a1d2f1b · outbound

This paper cites DecodingTrust: A comprehensive assessment of trustworthiness in GPT models.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection DecodingTrust: A comprehensive assessment of trustworthiness in GPT models

Reference 33

Resolution: verified fuzzy; raw_fallback, observed 2026-05-17T18:02:02.749703Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation 1bac7cee-21ce-49aa-9b01-83122f3bc2c9 · outbound

This paper cites Emoji Attack: Enhancing Jailbreak Attacks Against Judge LLM Detection.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection Emoji Attack: Enhancing Jailbreak Attacks Against Judge LLM Detection

Reference 34

Resolution: metadata mismatch; arxiv_id, observed 2026-05-11T09:41:00.448352Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation 44551253-064f-4a92-8988-1daf560d2621 · outbound

This paper cites Certified Data Removal from Machine Learning Models.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection Certified Data Removal from Machine Learning Models

Reference 35

Resolution: verified exact; arxiv_id, observed 2026-05-11T09:41:00.413001Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation 48367eca-1cbc-4b0f-83cd-540aaad00375 · outbound

This paper cites Detecting Backdoor Attacks on Deep Neural Networks by Activation Clustering.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection Detecting Backdoor Attacks on Deep Neural Networks by Activation Clustering

Reference 36

Resolution: verified exact; arxiv_id, observed 2026-05-11T09:41:00.428312Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Observation a51664cd-4d30-4cbf-836f-543a326db8cf · outbound

This paper cites Extracting training data from large language models.

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection Extracting training data from large language models

Reference 37

Resolution: verified fuzzy; raw_fallback, observed 2026-05-17T18:02:02.768647Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

Pith citing papers

No inbound Pith citation observations are available.