Paper Citation Record · LEDGER

CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning

As of 8 August 2026, this Paper Citation Record lists 24 of 24 outbound references and 3 inbound Pith citation observations for arXiv:2505.16559.

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

pith.paper-citation-record.v1
2505.16559 v1

Coverage vector

measured 24 of 24 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-08-07T15:02:35.374322Z

measured 27 of 27 standing notices

One-hop event checks from named stored sources.

Source: scholarly_work_events, retraction_status_cache, observed 2026-08-08T06:32:00.761636+00:00

measured 3 of 3 inbound itemization

Pith citing papers itemized under the disclosed page cap.

Source: paper_references, paper_reference_links, observed 2026-08-01T07:46:24.901516Z

measured 0 of 1 external citation measurements

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

Source: arxiv_reference, observed 2026-06-30T21:05:04.066789Z

Reference resolution

24 of 24 outbound references displayed

verified exact0
verified fuzzy15
unresolved8
parse uncertain0
malformed identifier1
metadata mismatch0

External citation measurements

No source-named external measurement is stored.

Outbound references

Observation 47eac5f4-625d-49d8-a030-1554e5ee909a · outbound

This paper cites Guidelines: • The answer NA means that the abstract and introduction do not include the claims made in the paper.

CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning Guidelines: • The answer NA means that the abstract and introduction do not include the claims made in the paper

Reference 1

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:02:37.960991Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation fc8b35b5-a7d2-4b17-9f04-d051e690e0b7 · outbound

This paper cites Guidelines: • The answer NA means that the paper has no limitation while the answer No means that the paper has limitations, but those are not discussed in the paper.

CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning Guidelines: • The answer NA means that the paper has no limitation while the answer No means that the paper has limitations, but those are not discussed in the paper

Reference 2

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:02:37.811234Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 8990bf3b-8501-47b8-8682-5589519135b9 · outbound

This paper cites Guidelines: • The answer NA means that the paper does not include theoretical results.

CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning Guidelines: • The answer NA means that the paper does not include theoretical results

Reference 3

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:02:37.684058Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation e7f5e475-6ba0-425f-a543-7848c54b26d0 · outbound

This paper cites LoRA: Low-Rank Adaptation of Large Language Models.

CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning LoRA: Low-Rank Adaptation of Large Language Models

Reference 4

Resolution: unresolved; no resolver link, observed 2026-08-07T15:02:33.819765Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation bb4026ff-ba31-4594-bd31-a1b658802edc · outbound

This paper cites Open Problems in Technical AI Governance.

CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning Open Problems in Technical AI Governance

Reference 5

Resolution: unresolved; no resolver link, observed 2026-08-07T15:02:33.948122Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e78ab843-a81a-4a19-b7b2-a8e195c0841d · outbound

This paper cites Evaluating Defences against Unsafe Feedback in RLHF.

CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning Evaluating Defences against Unsafe Feedback in RLHF

Reference 6

Resolution: unresolved; no resolver link, observed 2026-08-07T15:02:33.988579Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1cbd90fe-224d-45dd-bbd2-4e10f50e66a8 · outbound

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

CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning Gemma 2: Improving Open Language Models at a Practical Size

Reference 7

Resolution: unresolved; no resolver link, observed 2026-08-07T15:02:34.058916Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 84298f1d-7dff-4100-973e-0df839cc5d04 · outbound

This paper cites Qwen2 Technical Report.

CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning Qwen2 Technical Report

Reference 8

Resolution: malformed identifier; no resolver link, observed 2026-08-07T15:02:34.151266Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1b62cd71-aa38-417c-8a8f-534c1df27f3b · outbound

This paper cites Guidelines: • The answer NA means that the authors have not reviewed the NeurIPS Code of Ethics.

CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning Guidelines: • The answer NA means that the authors have not reviewed the NeurIPS Code of Ethics

Reference 9

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:02:36.772259Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 3c4dc9c7-5a88-4f59-94e4-b5ca76daa4e4 · outbound

This paper cites Guidelines: • The answer NA means that there is no societal impact of the work performed.

CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning Guidelines: • The answer NA means that there is no societal impact of the work performed

Reference 10

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:02:36.563700Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation f2a6db98-d21d-4027-8f4c-7070736d2c28 · outbound

This paper cites Guidelines: • The answer NA means that the paper poses no such risks.

CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning Guidelines: • The answer NA means that the paper poses no such risks

Reference 11

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:02:36.459060Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 12932260-186e-4354-b366-3a7235807a36 · outbound

This paper cites A more detailed version is in Appendix A.

CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning A more detailed version is in Appendix A

Reference 12

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:02:37.599162Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 14673d80-893b-4d0a-be1e-be9012348322 · outbound

This paper cites Guidelines: • The answer NA means that paper does not include experiments requiring code.

CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning Guidelines: • The answer NA means that paper does not include experiments requiring code

Reference 13

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:02:37.397411Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 8e63675a-aed4-41c8-b6a4-b4e44a93cbdd · outbound

This paper cites A more detailed version is in Appendix A.

CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning A more detailed version is in Appendix A

Reference 14

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:02:37.273431Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation ac28eb24-915c-4f3d-a34d-208d71f0f903 · outbound

This paper cites Guidelines: • The answer NA means that the paper does not include experiments.

CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning Guidelines: • The answer NA means that the paper does not include experiments

Reference 15

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:02:37.096302Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 5e9f117b-10e3-4ac0-90f9-5e76ebbce63e · outbound

This paper cites Guidelines: • The answer NA means that the paper does not include experiments.

CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning Guidelines: • The answer NA means that the paper does not include experiments

Reference 16

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:02:36.899419Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation c513294f-8e28-4519-9b52-c30b3e4d6cf9 · outbound

This paper cites Guidelines: • The answer NA means that the paper does not use existing assets.

CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning Guidelines: • The answer NA means that the paper does not use existing assets

Reference 20

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:02:36.359686Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation c9633bff-69c2-4211-836d-22008d98502b · outbound

This paper cites Guidelines: • The answer NA means that the paper does not release new assets.

CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning Guidelines: • The answer NA means that the paper does not release new assets

Reference 21

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:02:36.094759Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 8f2e6eff-983e-453b-9d64-b5766ce70ee3 · outbound

This paper cites Guidelines: • The answer NA means that the paper does not involve crowdsourcing nor research with human subjects.

CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning Guidelines: • The answer NA means that the paper does not involve crowdsourcing nor research with human subjects

Reference 22

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:02:35.901466Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 46dec3ee-c89c-461d-82a1-3ee33660ee97 · outbound

This paper cites Guidelines: • The answer NA means that the paper does not involve crowdsourcing nor research with human subjects.

Reference 23

Resolution: unresolved; no resolver link, observed 2026-08-07T15:02:35.319220Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b1210ffd-39ff-49eb-86d1-571cc18f6a94 · outbound

This paper cites Answer: [NA] Justification: This research does not incorporate LLMs as any important, original, or non- standard components in its core methodology development.

CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning Answer: [NA] Justification: This research does not incorporate LLMs as any important, original, or non- standard components in its core methodology development

Reference 24

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:02:35.722553Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 945327e5-42c1-4dcd-9ef0-14fb7d36d7b5 · outbound

This paper cites Toward Secure Tuning: Mitigating Security Risks from Instruction Fine-Tuning.

CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning Toward Secure Tuning: Mitigating Security Risks from Instruction Fine-Tuning

Reference 2023

Resolution: unresolved; no resolver link, observed 2026-08-07T15:02:33.690065Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a033fe85-b3d2-477b-b4a3-7b9377c2a16a · outbound

This paper cites Do as I do (Safely): Mitigating Task-Specific Fine-tuning Risks in Large Language Models.

CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning Do as I do (Safely): Mitigating Task-Specific Fine-tuning Risks in Large Language Models

Reference 2024

Resolution: unresolved; no resolver link, observed 2026-08-07T15:02:33.756311Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4257c855-f99b-4d55-ae36-2d0b4c91a86b · outbound

This paper cites Safety-Tuned LLaMAs: Lessons From Improving the Safety of Large Language Models that Follow Instructions.

CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning Safety-Tuned LLaMAs: Lessons From Improving the Safety of Large Language Models that Follow Instructions

Reference 2025

Resolution: unresolved; no resolver link, observed 2026-08-07T15:02:33.672549Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Pith citing papers

Observation 117067a0-9e10-4a3a-90b8-f3425c596817 · inbound

Token Buncher: Shielding LLMs from Harmful Reinforcement Learning Fine-Tuning cites this paper.

Token Buncher: Shielding LLMs from Harmful Reinforcement Learning Fine-Tuning CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning

Reference 61

Resolution: verified exact; arxiv_id, observed 2026-05-18T20:41:50.461365Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 8367f1ff-3e44-4ffa-85af-482337dac97c · inbound

One Step to the Side: Why Defenses Against Malicious Finetuning Fail Under Adaptive Adversaries cites this paper.

One Step to the Side: Why Defenses Against Malicious Finetuning Fail Under Adaptive Adversaries CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning

Reference 46

Resolution: verified exact; arxiv_id, observed 2026-06-30T21:05:04.068421Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 39c89248-31c4-49ef-b10c-8831b14fef65 · inbound

Emergent Misalignment Recruits a Pre-existing Persona Subspace cites this paper.

Emergent Misalignment Recruits a Pre-existing Persona Subspace CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning

Reference 236

Resolution: unresolved; no resolver link, observed 2026-08-01T07:46:24.901516Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.