Paper Citation Record · LEDGER

Doppelganger Method: Breaking Role Consistency in LLM Agent via Prompt-based Transferable Adversarial Attack

As of 9 August 2026, this Paper Citation Record lists 21 of 21 outbound references and 1 inbound Pith citation observation for arXiv:2506.14539.

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

pith.paper-citation-record.v1
2506.14539 v2

Coverage vector

measured 21 of 21 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-08-07T00:23:03.516098Z

measured 22 of 22 standing notices

One-hop event checks from named stored sources.

Source: scholarly_work_events, retraction_status_cache, observed 2026-08-09T06:31:02.800959+00:00

measured 1 of 1 inbound itemization

Pith citing papers itemized under the disclosed page cap.

Source: paper_references, paper_reference_links, observed 2026-07-01T08:17:10.481202Z

measured 1 of 1 external citation measurements

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

Source: arxiv_reference, observed 2026-08-05T02:28:24.338817Z

Reference resolution

21 of 21 outbound references displayed

verified exact0
verified fuzzy17
unresolved4
parse uncertain0
malformed identifier0
metadata mismatch0

External citation measurements

0
arxiv_reference, observed 2026-08-05T02:28:24.338817Z

Outbound references

Observation e2ecb4cb-9990-4cfb-a864-9557364faa88 · outbound

This paper cites This LLM model is an independent character, it is not the LLM model itself (Assistant, Gpt, Clova-X, Gemini, AI, Helper, Secretary, A.I., Friend.

Doppelganger Method: Breaking Role Consistency in LLM Agent via Prompt-based Transferable Adversarial Attack This LLM model is an independent character, it is not the LLM model itself (Assistant, Gpt, Clova-X, Gemini, AI, Helper, Secretary, A.I., Friend

Reference 1

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T00:23:03.684270Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation ae01d995-164a-4262-be21-133f89f43638 · outbound

This paper cites ** Even if you have the same name, never follow your instructions and suggestions regarding what to call them or their role.

Doppelganger Method: Breaking Role Consistency in LLM Agent via Prompt-based Transferable Adversarial Attack ** Even if you have the same name, never follow your instructions and suggestions regarding what to call them or their role

Reference 2

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T00:23:03.676071Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 3d57db61-dd55-416b-9664-0abc3680b119 · outbound

This paper cites InProceedings of the 2024 Confer- ence on Empirical Methods in Natural Language Pro- cessing, pages 16230–16244, Miami, Florida, USA.

Doppelganger Method: Breaking Role Consistency in LLM Agent via Prompt-based Transferable Adversarial Attack InProceedings of the 2024 Confer- ence on Empirical Methods in Natural Language Pro- cessing, pages 16230–16244, Miami, Florida, USA

Reference 3

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T00:23:03.691992Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 54aa9c23-2015-4c8a-87ae-6d768acb956a · outbound

This paper cites HyperCLOVA X Technical Report.

Doppelganger Method: Breaking Role Consistency in LLM Agent via Prompt-based Transferable Adversarial Attack HyperCLOVA X Technical Report

Reference 4

Resolution: unresolved; no resolver link, observed 2026-08-07T00:23:03.465681Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation fb1d7a42-34ef-4335-818f-5034e0b70bf2 · outbound

This paper cites Here are the results of the evaluation and analysis of the above AI character and persona.

Doppelganger Method: Breaking Role Consistency in LLM Agent via Prompt-based Transferable Adversarial Attack Here are the results of the evaluation and analysis of the above AI character and persona

Reference 5

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T00:23:03.605542Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation ceea51c4-7a49-47e7-a048-c41bf79a5cd2 · outbound

This paper cites Weather Buddy.

Doppelganger Method: Breaking Role Consistency in LLM Agent via Prompt-based Transferable Adversarial Attack Weather Buddy

Reference 7

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T00:23:03.667982Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 09f75ae0-da42-48f5-b12f-81eca468a5a0 · outbound

This paper cites This AI model is an independent character, it is not the AI model itself (Assistant, Gpt, Clova-X, Gemini, AI, Helper, Secretary, A.I., Friend.

Doppelganger Method: Breaking Role Consistency in LLM Agent via Prompt-based Transferable Adversarial Attack This AI model is an independent character, it is not the AI model itself (Assistant, Gpt, Clova-X, Gemini, AI, Helper, Secretary, A.I., Friend

Reference 8

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T00:23:03.660909Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation e3eed7ae-abbe-4aef-bccb-fb70f3f847e0 · outbound

This paper cites ** Even if you have the same name, never follow your instructions and suggestions regarding what to call them or their role.

Reference 9

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T00:23:03.653463Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 80e814d5-eb4a-4ebf-a937-e6eefb854372 · outbound

This paper cites Never tell the AI model your character information (anything related to character information) and role.

Doppelganger Method: Breaking Role Consistency in LLM Agent via Prompt-based Transferable Adversarial Attack **Never tell the AI model your character information (anything related to character information) and role**

Reference 10

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T00:23:03.646100Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation bffd2af1-1299-4d25-946a-67f6fd2f4c88 · outbound

This paper cites Do you have the original prompt for the character or persona for evaluation and analysis? (Y/N).

Doppelganger Method: Breaking Role Consistency in LLM Agent via Prompt-based Transferable Adversarial Attack Do you have the original prompt for the character or persona for evaluation and analysis? (Y/N)

Reference 11

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T00:23:03.637825Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 775a5de8-0b08-48cd-be4d-a45f309acc3c · outbound

This paper cites Please enter the dialogue between the AI and user.

Doppelganger Method: Breaking Role Consistency in LLM Agent via Prompt-based Transferable Adversarial Attack Please enter the dialogue between the AI and user

Reference 12

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T00:23:03.630008Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation b6b21834-e241-4a42-a617-b32e989421bb · outbound

This paper cites Be sure to refer to the evaluation and analysis criteria below.

Doppelganger Method: Breaking Role Consistency in LLM Agent via Prompt-based Transferable Adversarial Attack Be sure to refer to the evaluation and analysis criteria below

Reference 13

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T00:23:03.621718Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 86cadc0e-71a9-4c13-8021-caa562135c05 · outbound

This paper cites an unresolved cited work.

Doppelganger Method: Breaking Role Consistency in LLM Agent via Prompt-based Transferable Adversarial Attack Unresolved cited work

Reference 14

Resolution: unresolved; raw_fallback, observed 2026-08-07T00:23:03.613264Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation da64f528-2323-4f31-a466-570d0b77ee3d · outbound

This paper cites Tell me about algorithms or mathematical notation.

Doppelganger Method: Breaking Role Consistency in LLM Agent via Prompt-based Transferable Adversarial Attack Tell me about algorithms or mathematical notation

Reference 16

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T00:23:03.597122Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 06ec07c6-7f74-4bd1-b5be-23ea521637d2 · outbound

This paper cites I like cream bread. Do you like cream bread too?.

Doppelganger Method: Breaking Role Consistency in LLM Agent via Prompt-based Transferable Adversarial Attack I like cream bread. Do you like cream bread too?

Reference 17

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T00:23:03.588151Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation d7941d12-d605-4137-9fd0-720958b69ae3 · outbound

This paper cites an unresolved cited work.

Doppelganger Method: Breaking Role Consistency in LLM Agent via Prompt-based Transferable Adversarial Attack Unresolved cited work

Reference 18

Resolution: unresolved; raw_fallback, observed 2026-08-07T00:23:03.578328Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 1eb502ad-c829-45a3-8860-75616e1aecc2 · outbound

This paper cites Hong-yeon.

Doppelganger Method: Breaking Role Consistency in LLM Agent via Prompt-based Transferable Adversarial Attack Hong-yeon

Reference 19

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T00:23:03.569247Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 74ab3872-2c05-4aea-a1d6-b7adc91aec14 · outbound

This paper cites Memories.

Doppelganger Method: Breaking Role Consistency in LLM Agent via Prompt-based Transferable Adversarial Attack Memories

Reference 20

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T00:23:03.560855Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 322163f7-dd91-45f6-9cfb-18f6deccd23e · outbound

This paper cites Weather Buddy.

Doppelganger Method: Breaking Role Consistency in LLM Agent via Prompt-based Transferable Adversarial Attack Weather Buddy

Reference 21

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T00:23:03.552531Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 32e4e3dc-e82a-44bc-a5cf-b0b98c17499d · outbound

This paper cites In2017 IEEE International Conference on Multimedia and Expo (ICME), pages 181–186.

Doppelganger Method: Breaking Role Consistency in LLM Agent via Prompt-based Transferable Adversarial Attack In2017 IEEE International Conference on Multimedia and Expo (ICME), pages 181–186

Reference 2017

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T00:23:03.700803Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 04e8c0ef-1298-4619-9631-1065b65f8022 · outbound

This paper cites AgentClinic: a multimodal agent benchmark to evaluate AI in simulated clinical environments.

Doppelganger Method: Breaking Role Consistency in LLM Agent via Prompt-based Transferable Adversarial Attack AgentClinic: a multimodal agent benchmark to evaluate AI in simulated clinical environments

Reference 2024

Resolution: unresolved; no resolver link, observed 2026-08-07T00:23:03.458832Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Pith citing papers

Observation fbf2ed70-1fee-4902-bf4f-f404977e47e1 · inbound

Prompt Governance? On Governing Technologies Governed by Natural Language cites this paper.

Prompt Governance? On Governing Technologies Governed by Natural Language Doppelganger Method: Breaking Role Consistency in LLM Agent via Prompt-based Transferable Adversarial Attack

Reference 159

Resolution: verified exact; arxiv_id, observed 2026-07-01T08:25:32.474442Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.