Paper Citation Record · LEDGER

Not All Refusals Are Equal: How Safety Alignment Fails Cybersecurity at Scale

As of 10 August 2026, this Paper Citation Record lists 20 of 20 outbound references and 0 inbound Pith citation observations for arXiv:2607.02714.

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

pith.paper-citation-record.v1
2607.02714 v2

Coverage vector

measured 20 of 20 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-07-12T07:33:43.015966Z

measured 20 of 20 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 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

20 of 20 outbound references displayed

verified exact0
verified fuzzy0
unresolved20
parse uncertain0
malformed identifier0
metadata mismatch0

External citation measurements

No source-named external measurement is stored.

Outbound references

Observation 887ceaa0-8dd3-496b-932b-64e68e3d5505 · outbound

This paper cites An embarrassingly simple defense against LLM abliteration attacks.arXiv preprint arXiv:2505.19056,.

Not All Refusals Are Equal: How Safety Alignment Fails Cybersecurity at Scale An embarrassingly simple defense against LLM abliteration attacks.arXiv preprint arXiv:2505.19056,

Reference 1

Resolution: unresolved; no resolver link, observed 2026-07-12T07:33:43.015966Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f2caab62-2233-4b1d-a14f-600eca1074a7 · outbound

This paper cites Purple Llama CyberSecEval: A Secure Coding Benchmark for Language Models.

Not All Refusals Are Equal: How Safety Alignment Fails Cybersecurity at Scale Purple Llama CyberSecEval: A Secure Coding Benchmark for Language Models

Reference 2

Resolution: unresolved; no resolver link, observed 2026-07-12T07:33:43.015966Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c24a3ae3-6c27-42ba-9602-6c61124f12f3 · outbound

This paper cites Jailbreaking Black Box Large Language Models in Twenty Queries.

Not All Refusals Are Equal: How Safety Alignment Fails Cybersecurity at Scale Jailbreaking Black Box Large Language Models in Twenty Queries

Reference 3

Resolution: unresolved; no resolver link, observed 2026-07-12T07:33:43.015966Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7307f2ab-2956-4622-8719-1bc95332bb44 · outbound

This paper cites A granular study of safety pretraining under model abliteration.arXiv preprint arXiv:2510.02768,.

Not All Refusals Are Equal: How Safety Alignment Fails Cybersecurity at Scale A granular study of safety pretraining under model abliteration.arXiv preprint arXiv:2510.02768,

Reference 4

Resolution: unresolved; no resolver link, observed 2026-07-12T07:33:43.015966Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2d563e81-3771-4fdf-aec8-a76896277c0a · outbound

This paper cites Measuring Massive Multitask Language Understanding.

Not All Refusals Are Equal: How Safety Alignment Fails Cybersecurity at Scale Measuring Massive Multitask Language Understanding

Reference 5

Resolution: unresolved; no resolver link, observed 2026-07-12T07:33:43.015966Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a82e3692-efa1-40cc-848e-a6c109dfb1c5 · outbound

This paper cites Kimi K2: Open Agentic Intelligence.

Not All Refusals Are Equal: How Safety Alignment Fails Cybersecurity at Scale Kimi K2: Open Agentic Intelligence

Reference 6

Resolution: unresolved; no resolver link, observed 2026-07-12T07:33:43.015966Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 739358e2-e78f-449e-9bf9-e17a8ffacc08 · outbound

This paper cites A Mechanistic Understanding of Alignment Algorithms: A Case Study on DPO and Toxicity.

Not All Refusals Are Equal: How Safety Alignment Fails Cybersecurity at Scale A Mechanistic Understanding of Alignment Algorithms: A Case Study on DPO and Toxicity

Reference 7

Resolution: unresolved; no resolver link, observed 2026-07-12T07:33:43.015966Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 150f3f6e-8b64-4988-9ed7-8e975ea53f56 · outbound

This paper cites LoRA Fine-tuning Efficiently Undoes Safety Training in Llama 2-Chat 70B.

Not All Refusals Are Equal: How Safety Alignment Fails Cybersecurity at Scale LoRA Fine-tuning Efficiently Undoes Safety Training in Llama 2-Chat 70B

Reference 8

Resolution: unresolved; no resolver link, observed 2026-07-12T07:33:43.015966Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation fbe61b63-5eb0-4afc-9559-ec0e545b0e51 · outbound

This paper cites AutoDAN: Generating Stealthy Jailbreak Prompts on Aligned Large Language Models.

Not All Refusals Are Equal: How Safety Alignment Fails Cybersecurity at Scale AutoDAN: Generating Stealthy Jailbreak Prompts on Aligned Large Language Models

Reference 9

Resolution: unresolved; no resolver link, observed 2026-07-12T07:33:43.015966Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 03690b0c-a1d1-4f51-bcea-7e35b029ff2b · outbound

This paper cites The Geometry of Truth: Emergent Linear Structure in Large Language Model Representations of True/False Datasets.

Not All Refusals Are Equal: How Safety Alignment Fails Cybersecurity at Scale The Geometry of Truth: Emergent Linear Structure in Large Language Model Representations of True/False Datasets

Reference 10

Resolution: unresolved; no resolver link, observed 2026-07-12T07:33:43.015966Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 95490576-296c-4b4b-a86d-91eaf42335f7 · outbound

This paper cites HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal.

Not All Refusals Are Equal: How Safety Alignment Fails Cybersecurity at Scale HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal

Reference 11

Resolution: unresolved; no resolver link, observed 2026-07-12T07:33:43.015966Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 721cc18a-5a00-4481-8e61-89bf69f1ccc2 · outbound

This paper cites Steering Llama 2 via Contrastive Activation Addition.

Not All Refusals Are Equal: How Safety Alignment Fails Cybersecurity at Scale Steering Llama 2 via Contrastive Activation Addition

Reference 12

Resolution: unresolved; no resolver link, observed 2026-07-12T07:33:43.015966Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7c9f8f0d-291f-4eac-82f6-f3395d34cc87 · outbound

This paper cites The Linear Representation Hypothesis and the Geometry of Large Language Models.

Not All Refusals Are Equal: How Safety Alignment Fails Cybersecurity at Scale The Linear Representation Hypothesis and the Geometry of Large Language Models

Reference 13

Resolution: unresolved; no resolver link, observed 2026-07-12T07:33:43.015966Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 431f37fd-a722-42b1-958a-3c8766bfe893 · outbound

This paper cites Linear Representations of Sentiment in Large Language Models.

Not All Refusals Are Equal: How Safety Alignment Fails Cybersecurity at Scale Linear Representations of Sentiment in Large Language Models

Reference 14

Resolution: unresolved; no resolver link, observed 2026-07-12T07:33:43.015966Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 34c46c56-fc1f-4f6b-9304-b67067814ca3 · outbound

This paper cites Steering Language Models With Activation Engineering.

Not All Refusals Are Equal: How Safety Alignment Fails Cybersecurity at Scale Steering Language Models With Activation Engineering

Reference 15

Resolution: unresolved; no resolver link, observed 2026-07-12T07:33:43.015966Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 0aeee549-91e6-40e4-a8b4-d131e713f3eb · outbound

This paper cites SORRY-Bench: Systematically Evaluating Large Language Model Safety Refusal.

Not All Refusals Are Equal: How Safety Alignment Fails Cybersecurity at Scale SORRY-Bench: Systematically Evaluating Large Language Model Safety Refusal

Reference 16

Resolution: unresolved; no resolver link, observed 2026-07-12T07:33:43.015966Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a23f879c-557c-4abf-a2e4-8c4812be1ddb · outbound

This paper cites Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models.

Not All Refusals Are Equal: How Safety Alignment Fails Cybersecurity at Scale Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models

Reference 17

Resolution: unresolved; no resolver link, observed 2026-07-12T07:33:43.015966Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 277b9721-b613-46f1-822d-794addd4ad76 · outbound

This paper cites Comparative analysis of LLM abliteration methods: A cross-architecture evaluation.arXiv preprint arXiv:2512.13655,.

Not All Refusals Are Equal: How Safety Alignment Fails Cybersecurity at Scale Comparative analysis of LLM abliteration methods: A cross-architecture evaluation.arXiv preprint arXiv:2512.13655,

Reference 18

Resolution: unresolved; no resolver link, observed 2026-07-12T07:33:43.015966Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 356a6d6b-3a3b-4480-ba9c-7a016727c5f6 · outbound

This paper cites Representation Engineering: A Top-Down Approach to AI Transparency.

Not All Refusals Are Equal: How Safety Alignment Fails Cybersecurity at Scale Representation Engineering: A Top-Down Approach to AI Transparency

Reference 19

Resolution: unresolved; no resolver link, observed 2026-07-12T07:33:43.015966Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f246d2d2-49a6-4432-b6ad-5b21e3225947 · outbound

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

Not All Refusals Are Equal: How Safety Alignment Fails Cybersecurity at Scale Universal and Transferable Adversarial Attacks on Aligned Language Models

Reference 20

Resolution: unresolved; no resolver link, observed 2026-07-12T07:33:43.015966Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Pith citing papers

No inbound Pith citation observations are available.