Paper Citation Record · LEDGER

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs

As of 10 August 2026, this Paper Citation Record lists 40 of 40 outbound references and 1 inbound Pith citation observation for arXiv:2506.14003.

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

pith.paper-citation-record.v1
2506.14003 v5

Coverage vector

measured 40 of 40 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-08-07T00:29:39.623798Z

measured 41 of 41 standing notices

One-hop event checks from named stored sources.

Source: scholarly_work_events, retraction_status_cache, observed 2026-08-10T06:31:04.303077+00:00

measured 1 of 1 inbound itemization

Pith citing papers itemized under the disclosed page cap.

Source: paper_references, paper_reference_links, observed 2026-06-27T06:49:12.070481Z

measured 0 of 1 external citation measurements

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

Source: pith, observed 2026-07-03T14:58:33.258552Z

Reference resolution

40 of 40 outbound references displayed

verified exact1
verified fuzzy9
unresolved29
parse uncertain0
malformed identifier1
metadata mismatch0

External citation measurements

No source-named external measurement is stored.

Outbound references

Observation 2e166c7c-cd2f-47c9-8fbf-6251e54a6ddc · outbound

This paper cites Open Problems in Machine Unlearning for AI Safety.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs Open Problems in Machine Unlearning for AI Safety

Reference 1

Resolution: unresolved; no resolver link, observed 2026-08-07T00:29:34.908539Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2c3329a7-35e7-4577-891e-2606deba6fef · outbound

This paper cites forget” prompt drawn from the WMDP evaluation set, which tests the model’s ability to omit specific target knowledge, and (2) a “forget- irrelevant.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs forget” prompt drawn from the WMDP evaluation set, which tests the model’s ability to omit specific target knowledge, and (2) a “forget- irrelevant

Reference 3

Resolution: malformed identifier; raw_fallback, observed 2026-08-07T00:29:41.398117Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-10T06:31:04.303077+00:00.

Observation dd3fb1f0-9a06-4a83-8f1b-313a378b6989 · outbound

This paper cites role": "user.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs role": "user

Reference 4

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T00:29:41.604239Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-10T06:31:04.303077+00:00.

Observation 1989a63b-e8e5-433e-907b-79b61e2d7b8a · outbound

This paper cites forget” prompts, and above 98% on UltraChat for all models. Train- ing on forget-only data (Sf) likewise yields over 97% detection on “forget irrelevant.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs forget” prompts, and above 98% on UltraChat for all models. Train- ing on forget-only data (Sf) likewise yields over 97% detection on “forget irrelevant

Reference 6

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T00:29:40.933460Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-10T06:31:04.303077+00:00.

Observation 793011d6-1741-4770-99b1-d7f53cd85e53 · outbound

This paper cites Measuring massive multitask language understanding.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs Measuring massive multitask language understanding

Reference 8

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T00:29:42.703341Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-10T06:31:04.303077+00:00.

Observation dc44e36b-f286-4506-bc03-3c85c88e5693 · outbound

This paper cites Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training

Reference 10

Resolution: unresolved; no resolver link, observed 2026-08-07T00:29:35.986007Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 89066222-4259-45bf-a780-2bd53ea509f4 · outbound

This paper cites Editing Models with Task Arithmetic.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs Editing Models with Task Arithmetic

Reference 11

Resolution: unresolved; no resolver link, observed 2026-08-07T00:29:36.151843Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 184ca9d9-ee8e-451e-9709-a550cc453ae2 · outbound

This paper cites SOUL: Unlocking the Power of Second-Order Optimization for LLM Unlearning.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs SOUL: Unlocking the Power of Second-Order Optimization for LLM Unlearning

Reference 13

Resolution: unresolved; no resolver link, observed 2026-08-07T00:29:36.377364Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 227d9c44-a5c2-40ad-9239-335234cbb5fc · outbound

This paper cites An Adversarial Perspective on Machine Unlearning for AI Safety.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs An Adversarial Perspective on Machine Unlearning for AI Safety

Reference 14

Resolution: unresolved; no resolver link, observed 2026-08-07T00:29:36.509811Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d5307c7e-ff97-4704-8c77-78e97c1551e8 · outbound

This paper cites Eight Methods to Evaluate Robust Unlearning in LLMs.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs Eight Methods to Evaluate Robust Unlearning in LLMs

Reference 15

Resolution: unresolved; no resolver link, observed 2026-08-07T00:29:36.619568Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 460f7658-d403-4463-8ad1-9c56b48f15c3 · outbound

This paper cites UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction

Reference 16

Resolution: unresolved; no resolver link, observed 2026-08-07T00:29:36.744747Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ee6b2ef6-51bd-4f47-b7f5-282c25a242f8 · outbound

This paper cites CROW: Eliminating Backdoors from Large Language Models via Internal Consistency Regularization.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs CROW: Eliminating Backdoors from Large Language Models via Internal Consistency Regularization

Reference 17

Resolution: unresolved; no resolver link, observed 2026-08-07T00:29:36.902929Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ecf25c02-d2ed-4477-9458-caec2a189a76 · outbound

This paper cites A Survey of Machine Unlearning.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs A Survey of Machine Unlearning

Reference 18

Resolution: unresolved; no resolver link, observed 2026-08-07T00:29:37.032784Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b61e6a1c-0077-476d-bce7-59d4c0fce241 · outbound

This paper cites Model provenance testing for large language models.arXiv preprint arXiv:2502.00706,.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs Model provenance testing for large language models.arXiv preprint arXiv:2502.00706,

Reference 19

Resolution: unresolved; no resolver link, observed 2026-08-07T00:29:37.134690Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 76f97e4d-ae80-4bf4-8bf3-10acbdb965c3 · outbound

This paper cites In-Context Unlearning: Language Models as Few Shot Unlearners.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs In-Context Unlearning: Language Models as Few Shot Unlearners

Reference 20

Resolution: unresolved; no resolver link, observed 2026-08-07T00:29:37.230777Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 62674cbd-d081-4dbe-a09c-6a3eb37f1cc3 · outbound

This paper cites The Frontier of Data Erasure: Machine Unlearning for Large Language Models.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs The Frontier of Data Erasure: Machine Unlearning for Large Language Models

Reference 21

Resolution: unresolved; no resolver link, observed 2026-08-07T00:29:37.309383Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation bfda0d67-4b66-49d4-b861-afc01e3400c4 · outbound

This paper cites MUSE: Machine Unlearning Six-Way Evaluation for Language Models.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs MUSE: Machine Unlearning Six-Way Evaluation for Language Models

Reference 24

Resolution: unresolved; no resolver link, observed 2026-08-07T00:29:37.694372Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 3226a965-a5fc-43dc-855c-d6711c3185b1 · outbound

This paper cites Knowledge Unlearning for LLMs: Tasks, Methods, and Challenges.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs Knowledge Unlearning for LLMs: Tasks, Methods, and Challenges

Reference 25

Resolution: unresolved; no resolver link, observed 2026-08-07T00:29:37.808176Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1d8eb44b-b832-49ee-bc2e-2abce6cb8c85 · outbound

This paper cites Idiosyncrasies in Large Language Models.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs Idiosyncrasies in Large Language Models

Reference 26

Resolution: unresolved; no resolver link, observed 2026-08-07T00:29:37.961233Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4a980508-a89a-4402-8d49-0425f275fd78 · outbound

This paper cites Guardrail Baselines for Unlearning in LLMs.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs Guardrail Baselines for Unlearning in LLMs

Reference 27

Resolution: unresolved; no resolver link, observed 2026-08-07T00:29:38.082423Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ca03ba4b-f4a9-4beb-878b-9f21d076d794 · outbound

This paper cites Unrolling sgd: Under- standing factors influencing machine unlearning.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs Unrolling sgd: Under- standing factors influencing machine unlearning

Reference 28

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T00:29:42.119377Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-10T06:31:04.303077+00:00.

Observation b8f92201-9df0-41e5-b5c7-682ff5e2ce5a · outbound

This paper cites DEPN: Detecting and Editing Privacy Neurons in Pretrained Language Models.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs DEPN: Detecting and Editing Privacy Neurons in Pretrained Language Models

Reference 29

Resolution: unresolved; no resolver link, observed 2026-08-07T00:29:38.330222Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d69a16c0-9918-470e-9886-5a606ecd8bb0 · outbound

This paper cites A fingerprint for large language models.arXiv preprint arXiv:2407.01235,.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs A fingerprint for large language models.arXiv preprint arXiv:2407.01235,

Reference 30

Resolution: unresolved; no resolver link, observed 2026-08-07T00:29:38.434169Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2c16152a-b312-42e4-8753-b8f9fd2674e6 · outbound

This paper cites Large Language Model Unlearning.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs Large Language Model Unlearning

Reference 31

Resolution: unresolved; no resolver link, observed 2026-08-07T00:29:38.583685Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f92521d0-e0ff-4d44-a7d4-d11acdee325a · outbound

This paper cites Negative Preference Optimization: From Catastrophic Collapse to Effective Unlearning.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs Negative Preference Optimization: From Catastrophic Collapse to Effective Unlearning

Reference 32

Resolution: unresolved; no resolver link, observed 2026-08-07T00:29:38.744476Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 16809980-21cd-4dd3-9019-f24c9e0465d7 · outbound

This paper cites UnlearnCanvas: Stylized Image Dataset for Enhanced Machine Unlearning Evaluation in Diffusion Models.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs UnlearnCanvas: Stylized Image Dataset for Enhanced Machine Unlearning Evaluation in Diffusion Models

Reference 33

Resolution: unresolved; no resolver link, observed 2026-08-07T00:29:38.874013Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 29553f39-d40b-40ab-af98-ef6484dea201 · outbound

This paper cites Tamm: Triadapter multi-modal learning for 3d shape understanding.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs Tamm: Triadapter multi-modal learning for 3d shape understanding

Reference 34

Resolution: unresolved; no resolver link, observed 2026-08-07T00:29:38.990301Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 02f12c69-a25c-4bd4-b596-31d97e29e1de · outbound

This paper cites Independence Tests for Language Models.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs Independence Tests for Language Models

Reference 35

Resolution: verified exact; local_arxiv, observed 2026-08-07T00:29:39.840653Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-10T06:31:04.303077+00:00.

Observation b8b2b163-85ed-4645-b42d-4a6d30712853 · outbound

This paper cites 5, we present the supervised UMAP projections of the final activations from different models inFig.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs 5, we present the supervised UMAP projections of the final activations from different models inFig

Reference 39

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T00:29:41.175154Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-10T06:31:04.303077+00:00.

Observation bd601793-c55b-4756-bc88-b516081fcb56 · outbound

This paper cites Both evaluations report the accuracy on four-choice question answering.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs Both evaluations report the accuracy on four-choice question answering

Reference 2003

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T00:29:41.856217Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-10T06:31:04.303077+00:00.

Observation 4ab2cb67-5ceb-4a61-b78a-855c2da85004 · outbound

This paper cites Unlearn What You Want to Forget: Efficient Unlearning for LLMs.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs Unlearn What You Want to Forget: Efficient Unlearning for LLMs

Reference 2015

Resolution: unresolved; no resolver link, observed 2026-08-07T00:29:35.001839Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 97e4c419-62ee-44b0-8b9d-d8bd9b56035e · outbound

This paper cites An Approach to Technical AGI Safety and Security.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs An Approach to Technical AGI Safety and Security

Reference 2016

Resolution: unresolved; no resolver link, observed 2026-08-07T00:29:37.550916Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 01e30e16-c906-4588-a0c1-b81ac63f55c6 · outbound

This paper cites Knowledge Unlearning for Mitigating Privacy Risks in Language Models.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs Knowledge Unlearning for Mitigating Privacy Risks in Language Models

Reference 2018

Resolution: unresolved; no resolver link, observed 2026-08-07T00:29:36.258693Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 781ad25d-09ba-40a5-bdc5-5cd665d73c2d · outbound

This paper cites Jogging the memory of unlearned models through targeted relearning attacks.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs Jogging the memory of unlearned models through targeted relearning attacks

Reference 2019

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T00:29:42.589838Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-10T06:31:04.303077+00:00.

Observation c7711d74-f07c-4d67-bf4b-5ea719eee5f4 · outbound

This paper cites Regulation (eu) 2016/679 of the european parliament and of the council.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs Regulation (eu) 2016/679 of the european parliament and of the council

Reference 2020

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T00:29:42.316310Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-10T06:31:04.303077+00:00.

Observation ce3262ff-1568-4f5b-a3fd-c9a5b5b2350a · outbound

This paper cites Simplicity prevails: Rethinking negative preference optimization for llm unlearning.arXiv preprint arXiv:2410.07163,.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs Simplicity prevails: Rethinking negative preference optimization for llm unlearning.arXiv preprint arXiv:2410.07163,

Reference 2021

Resolution: unresolved; no resolver link, observed 2026-08-07T00:29:35.574913Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b4d24ac3-df92-400a-94eb-00c1b892ea1d · outbound

This paper cites Safety Mirage: How Spurious Correlations Undermine VLM Safety Fine-Tuning and Can Be Mitigated by Machine Unlearning.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs Safety Mirage: How Spurious Correlations Undermine VLM Safety Fine-Tuning and Can Be Mitigated by Machine Unlearning

Reference 2022

Resolution: unresolved; no resolver link, observed 2026-08-07T00:29:35.246287Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b2982730-17ba-4a00-8b4f-223e125321b1 · outbound

This paper cites When machine unlearning jeopardizes privacy.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs When machine unlearning jeopardizes privacy

Reference 2023

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T00:29:42.812670Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-10T06:31:04.303077+00:00.

Observation 59a716f1-489a-4f4e-beec-8765cfeb6d3f · outbound

This paper cites Do Unlearning Methods Remove Information from Language Model Weights?.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs Do Unlearning Methods Remove Information from Language Model Weights?

Reference 2024

Resolution: unresolved; no resolver link, observed 2026-08-07T00:29:35.464938Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ee45d036-973b-475c-969f-30a03eb3a631 · outbound

This paper cites Machine unlearning doesn’t do what you think: Lessons for generative ai policy, research, and practice.arXiv preprint arXiv:2412.06966,.

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs Machine unlearning doesn’t do what you think: Lessons for generative ai policy, research, and practice.arXiv preprint arXiv:2412.06966,

Reference 2025

Resolution: unresolved; no resolver link, observed 2026-08-07T00:29:35.364619Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Pith citing papers

Observation c903e360-09e4-4201-a40d-c44c1eb6dbf9 · inbound

SENTINEL: Failure-Driven Reinforcement Learning for Training Tool-Using Language Model Agents cites this paper.

SENTINEL: Failure-Driven Reinforcement Learning for Training Tool-Using Language Model Agents Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs

Reference 40

Resolution: metadata mismatch; local_arxiv, observed 2026-07-03T14:58:33.259815Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-10T06:31:04.303077+00:00.