Typed states for the displayed outbound observations.
Source: paper_references, paper_reference_links
Paper Citation Record · LEDGER
Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
As of 22 August 2026, this Paper Citation Record lists 0 of 0 outbound references and 67 inbound Pith citation observations for arXiv:2310.02949.
A citation records a reference. It does not transfer a finding from one paper to another.
Coverage vector
One-hop event checks from named stored sources.
Source: scholarly_work_events, retraction_status_cache, observed 2026-08-22T06:32:14.747728+00:00
Pith citing papers itemized under the disclosed page cap.
Source: paper_references, paper_reference_links, observed 2026-08-16T11:24:12.725601Z
A source-named dated measurement, never combined with another source.
Source: arxiv_reference, observed 2026-07-04T09:09:43.659817Z
Reference resolution
0 of 0 outbound references displayed
- verified exact0
- verified fuzzy0
- unresolved0
- parse uncertain0
- malformed identifier0
- metadata mismatch0
External citation measurements
No source-named external measurement is stored.
Outbound references
No outbound reference observations are available for this paper version.
Pith citing papers
Observation 52fff216-cd20-4ab6-bdac-3a7b0be0fd59 · inbound
LLM Agents can Autonomously Exploit One-day Vulnerabilities cites this paper.
LLM Agents can Autonomously Exploit One-day Vulnerabilities Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 21
- Resolution
- metadata mismatch
- arxiv_id, observed 2026-05-18T04:18:27.647836Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation d7a14b8b-4a05-474c-aec7-8cd21659b5b8 · inbound
Refusal in Language Models Is Mediated by a Single Direction cites this paper.
Refusal in Language Models Is Mediated by a Single Direction Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 202
- Resolution
- metadata mismatch
- arxiv_id, observed 2026-05-13T10:47:56.163266Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation b20987fe-9db7-4022-a645-8d8002f9c10b · inbound
Jailbreak Attacks and Defenses Against Large Language Models: A Survey cites this paper.
Jailbreak Attacks and Defenses Against Large Language Models: A Survey Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 103
- Resolution
- verified exact
- arxiv_id, observed 2026-05-15T02:20:44.444563Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation 1107c248-3eca-47d0-ada5-8b11aaa71381 · inbound
Learning to Ask: When LLM Agents Meet Unclear Instruction cites this paper.
Learning to Ask: When LLM Agents Meet Unclear Instruction Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 32
- Resolution
- verified exact
- arxiv_id, observed 2026-05-23T21:13:28.074522Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation e7154fa9-a96d-40ee-9d84-d5124b36211b · inbound
Harmful Fine-tuning Attacks and Defenses for Large Language Models: A Survey cites this paper.
Harmful Fine-tuning Attacks and Defenses for Large Language Models: A Survey Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 167
- Resolution
- verified exact
- arxiv_id, observed 2026-05-23T20:58:26.176221Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation 97539c8c-b6f4-4227-afd2-b0f9359bf3ae · inbound
Steering Language Model Refusal with Sparse Autoencoders cites this paper.
Steering Language Model Refusal with Sparse Autoencoders Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 55
- Resolution
- unresolved
- no resolver link, observed 2026-08-12T18:45:48.602638Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 2bbfcfad-e45c-48b8-8223-07a077794e2d · inbound
RV4Chatbot: Are Chatbots Allowed to Dream of Electric Sheep? cites this paper.
RV4Chatbot: Are Chatbots Allowed to Dream of Electric Sheep? Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 46
- Resolution
- unresolved
- no resolver link, observed 2026-08-12T15:18:19.480456Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 26b8fe06-b213-4a31-acf4-60fd49e7fbcf · inbound
Separate the Wheat from the Chaff: A Post-Hoc Approach to Safety Re-Alignment for Fine-Tuned Language Models cites this paper.
Separate the Wheat from the Chaff: A Post-Hoc Approach to Safety Re-Alignment for Fine-Tuned Language Models Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 44
- Resolution
- unresolved
- no resolver link, observed 2026-08-11T15:26:52.629530Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 5c404844-43ab-4d8a-8466-32a1ffca7b90 · inbound
NLSR: Neuron-Level Safety Realignment of Large Language Models Against Harmful Fine-Tuning cites this paper.
NLSR: Neuron-Level Safety Realignment of Large Language Models Against Harmful Fine-Tuning Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 35
- Resolution
- unresolved
- no resolver link, observed 2026-08-11T14:06:48.213034Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 4d06135b-4abe-45b6-9b06-1baf948d5a2f · inbound
Crabs: Consuming Resource via Auto-generation for LLM-DoS Attack under Black-box Settings cites this paper.
Crabs: Consuming Resource via Auto-generation for LLM-DoS Attack under Black-box Settings Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 10
- Resolution
- unresolved
- no resolver link, observed 2026-08-11T12:48:10.938110Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 303e0b88-975d-4769-a3c8-8d356b9950a2 · inbound
Safeguard Fine-Tuned LLMs Through Pre- and Post-Tuning Model Merging cites this paper.
Safeguard Fine-Tuned LLMs Through Pre- and Post-Tuning Model Merging Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 48
- Resolution
- unresolved
- no resolver link, observed 2026-08-11T00:18:57.322310Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 35f49797-2e6a-4e80-ad93-790aa4ed03df · inbound
Self-Instruct Few-Shot Jailbreaking: Decompose the Attack into Pattern and Behavior Learning cites this paper.
Self-Instruct Few-Shot Jailbreaking: Decompose the Attack into Pattern and Behavior Learning Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 45
- Resolution
- unresolved
- no resolver link, observed 2026-08-10T20:35:52.870948Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 34195a89-002c-4de1-8b08-2cfc602de11c · inbound
LLMs are Vulnerable to Malicious Prompts Disguised as Scientific Language cites this paper.
LLMs are Vulnerable to Malicious Prompts Disguised as Scientific Language Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 48
- Resolution
- unresolved
- no resolver link, observed 2026-08-10T15:26:44.965317Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 81b9a09f-61db-463b-8b6d-d0322382fd37 · inbound
Vulnerability Mitigation for Safety-Aligned Language Models via Debiasing cites this paper.
Vulnerability Mitigation for Safety-Aligned Language Models via Debiasing Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 51
- Resolution
- unresolved
- no resolver link, observed 2026-08-09T13:14:34.132795Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation da6e0ecc-79bb-47cc-80ab-863b8edc129c · inbound
Model Tampering Attacks Enable More Rigorous Evaluations of LLM Capabilities cites this paper.
Model Tampering Attacks Enable More Rigorous Evaluations of LLM Capabilities Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 79
- Resolution
- unresolved
- no resolver link, observed 2026-08-09T14:47:15.377658Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 673dce01-3784-422e-be3e-2367f20e51a7 · inbound
A Comprehensive Survey in LLM(-Agent) Full Stack Safety: Data, Training and Deployment cites this paper.
A Comprehensive Survey in LLM(-Agent) Full Stack Safety: Data, Training and Deployment Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 239
- Resolution
- unresolved
- no resolver link, observed 2026-08-16T11:24:12.725601Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 61ff8dc4-fde5-4294-86d1-12f67372e810 · inbound
Latent Adversarial Training Improves the Representation of Refusal cites this paper.
Latent Adversarial Training Improves the Representation of Refusal Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 13
- Resolution
- unresolved
- no resolver link, observed 2026-08-16T10:11:37.865408Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation b0439625-cd98-473e-864e-872bba7e7498 · inbound
Unlearning Sensitive Information in Multimodal LLMs: Benchmark and Attack-Defense Evaluation cites this paper.
Unlearning Sensitive Information in Multimodal LLMs: Benchmark and Attack-Defense Evaluation Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 55
- Resolution
- unresolved
- no resolver link, observed 2026-08-16T04:52:53.136494Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 5e371af3-f895-4cb0-9836-39478243d101 · inbound
Fight Fire with Fire: Defending Against Malicious RL Fine-Tuning via Reward Neutralization cites this paper.
Fight Fire with Fire: Defending Against Malicious RL Fine-Tuning via Reward Neutralization Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 26
- Resolution
- unresolved
- no resolver link, observed 2026-08-15T23:30:24.291975Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 2e6a2e86-1dfb-49eb-bca1-21e2781a527e · inbound
Vulnerability-Aware Alignment: Mitigating Uneven Forgetting in Harmful Fine-Tuning cites this paper.
Vulnerability-Aware Alignment: Mitigating Uneven Forgetting in Harmful Fine-Tuning Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 41
- Resolution
- unresolved
- no resolver link, observed 2026-08-07T10:58:32.316238Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 574b4f9d-c622-4a9a-873a-34bff906c65b · inbound
Fine-Tuning Lowers Safety and Disrupts Evaluation Consistency cites this paper.
Fine-Tuning Lowers Safety and Disrupts Evaluation Consistency Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 33
- Resolution
- unresolved
- no resolver link, observed 2026-08-06T23:33:41.022429Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 8423a4ca-4720-4e34-9c9a-ca8ba783decd · inbound
SoK: A Comprehensive Security Analysis of Jailbreak Resilience in GPT and DeepSeek Models cites this paper.
SoK: A Comprehensive Security Analysis of Jailbreak Resilience in GPT and DeepSeek Models Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 53
- Resolution
- unresolved
- no resolver link, observed 2026-08-06T23:20:57.826882Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 301aec01-640b-4ea0-af6f-624097a44f17 · inbound
Safe Pruning LoRA: Robust Distance-Guided Pruning for Safety Alignment in Adaptation of LLMs cites this paper.
Safe Pruning LoRA: Robust Distance-Guided Pruning for Safety Alignment in Adaptation of LLMs Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 37
- Resolution
- unresolved
- no resolver link, observed 2026-08-15T19:09:30.224427Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 916ad6f6-f012-41d2-8414-a9fbb17f21ee · inbound
Innocence in the Crossfire: Roles of Skip Connections in Jailbreaking Visual Language Models cites this paper.
Innocence in the Crossfire: Roles of Skip Connections in Jailbreaking Visual Language Models Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 41
- Resolution
- unresolved
- no resolver link, observed 2026-08-06T16:25:55.069605Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation b7964131-1802-4170-9105-c6f4ee58b793 · inbound
SDD: Self-Degraded Defense against Malicious Fine-tuning cites this paper.
SDD: Self-Degraded Defense against Malicious Fine-tuning Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 56
- Resolution
- unresolved
- no resolver link, observed 2026-08-15T17:55:14.032498Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation e7c2c932-e692-4384-ad68-e40230d919db · inbound
Layer-Wise Perturbations via Sparse Autoencoders for Adversarial Text Generation cites this paper.
Layer-Wise Perturbations via Sparse Autoencoders for Adversarial Text Generation Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 41
- Resolution
- unresolved
- no resolver link, observed 2026-08-05T20:31:34.960696Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation bcfcf42a-f78e-4e05-8033-fde1fe9220d5 · inbound
Consiglieres in the Shadow: Understanding the Use of Uncensored Large Language Models in Cybercrimes cites this paper.
Consiglieres in the Shadow: Understanding the Use of Uncensored Large Language Models in Cybercrimes Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 161
- Resolution
- unresolved
- no resolver link, observed 2026-08-15T17:24:59.641049Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 6c208afa-7b3a-4326-84bd-828646d8ed1c · inbound
S3LoRA: Safe Spectral Sharpness-Guided Pruning in Adaptation of Agent Planner cites this paper.
S3LoRA: Safe Spectral Sharpness-Guided Pruning in Adaptation of Agent Planner Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 37
- Resolution
- unresolved
- no resolver link, observed 2026-08-05T18:12:36.832306Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 5599eacc-8ae3-4e23-9f71-3f74990be721 · inbound
Turning the Spell Around: Lightweight Alignment Amplification via Rank-One Safety Injection cites this paper.
Turning the Spell Around: Lightweight Alignment Amplification via Rank-One Safety Injection Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 54
- Resolution
- unresolved
- no resolver link, observed 2026-08-05T14:57:12.949136Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation d1737b1d-a775-4801-93c3-1501881e726a · inbound
Between a Rock and a Hard Place: The Tension Between Ethical Reasoning and Safety Alignment in LLMs cites this paper.
Between a Rock and a Hard Place: The Tension Between Ethical Reasoning and Safety Alignment in LLMs Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 7
- Resolution
- verified exact
- arxiv_id, observed 2026-05-18T19:36:47.456783Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation 4b329b83-cbf4-4dc4-b7f5-9c8b8ee0aaee · inbound
Between a Rock and a Hard Place: The Tension Between Ethical Reasoning and Safety Alignment in LLMs cites this paper.
Between a Rock and a Hard Place: The Tension Between Ethical Reasoning and Safety Alignment in LLMs Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 7
- Resolution
- unresolved
- no resolver link, observed 2026-08-05T10:36:15.882098Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 2f46b1c3-f228-4b2d-b67f-0d8c055eba2c · inbound
Embedding Poisoning: Bypassing Safety Alignment via Embedding Semantic Shift cites this paper.
Embedding Poisoning: Bypassing Safety Alignment via Embedding Semantic Shift Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 40
- Resolution
- unresolved
- no resolver link, observed 2026-08-15T16:21:58.950836Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation a9274f2a-95f3-4387-aa55-e4ea777a042f · inbound
Paladin: Defending LLM-enabled Phishing Emails with a New Trigger-Tag Paradigm cites this paper.
Paladin: Defending LLM-enabled Phishing Emails with a New Trigger-Tag Paradigm Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 84
- Resolution
- unresolved
- no resolver link, observed 2026-08-04T22:33:28.772074Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation a3597325-4084-4808-804f-767418eaaee7 · inbound
Efficiency vs. Alignment: Investigating Safety and Fairness Risks in Parameter-Efficient Fine-Tuning of LLMs cites this paper.
Efficiency vs. Alignment: Investigating Safety and Fairness Risks in Parameter-Efficient Fine-Tuning of LLMs Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 46
- Resolution
- unresolved
- no resolver link, observed 2026-08-04T06:54:11.503514Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 30a2df69-b7c7-4aa0-bc5f-0db29504e29f · inbound
Robust Policy Optimization to Prevent Catastrophic Forgetting cites this paper.
Robust Policy Optimization to Prevent Catastrophic Forgetting Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 55
- Resolution
- metadata mismatch
- arxiv_id, observed 2026-05-16T05:37:24.200106Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation 539c423a-598c-47bd-a544-9f2cdb64d3e4 · inbound
The Art of (Mis)alignment: How Fine-Tuning Methods Effectively Misalign and Realign LLMs in Post-Training cites this paper.
The Art of (Mis)alignment: How Fine-Tuning Methods Effectively Misalign and Realign LLMs in Post-Training Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 62
- Resolution
- metadata mismatch
- arxiv_id, observed 2026-05-11T00:45:50.626971Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation 4a20819c-bd6d-4dac-97c7-d940b96d1f11 · inbound
Continual Safety Alignment via Gradient-Based Sample Selection cites this paper.
Continual Safety Alignment via Gradient-Based Sample Selection Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 7
- Resolution
- verified exact
- arxiv_id, observed 2026-05-10T07:16:54.723710Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation 3982e68c-4587-4aa8-b7ef-d5b29f6f0bb2 · inbound
Representation-Guided Parameter-Efficient LLM Unlearning cites this paper.
Representation-Guided Parameter-Efficient LLM Unlearning Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 42
- Resolution
- metadata mismatch
- arxiv_id, observed 2026-05-10T06:06:19.395425Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation ac357e15-b452-4ffa-9014-d6ea0e5def20 · inbound
A Systematic Survey of Security Threats and Defenses in LLM-Based AI Agents: A Layered Attack Surface Framework cites this paper.
A Systematic Survey of Security Threats and Defenses in LLM-Based AI Agents: A Layered Attack Surface Framework Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 42
- Resolution
- metadata mismatch
- arxiv_id, observed 2026-05-11T20:51:09.500341Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation 6c007f23-3612-4364-9134-03e5d47b5c66 · inbound
Safety Drift After Fine-Tuning: Evidence from High-Stakes Domains cites this paper.
Safety Drift After Fine-Tuning: Evidence from High-Stakes Domains Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 69
- Resolution
- verified exact
- arxiv_id, observed 2026-05-11T23:16:14.005373Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation 026f3de0-b29c-4319-8c5b-8b46131ae44a · inbound
When Safety Geometry Collapses: Fine-Tuning Vulnerabilities in Agentic Guard Models cites this paper.
When Safety Geometry Collapses: Fine-Tuning Vulnerabilities in Agentic Guard Models Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 21
- Resolution
- metadata mismatch
- arxiv_id, observed 2026-05-11T00:05:49.280147Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation 6b05f4bd-9405-42b9-bc68-ec7f5a25c467 · inbound
Few-Shot Truly Benign DPO Attack for Jailbreaking LLMs cites this paper.
Few-Shot Truly Benign DPO Attack for Jailbreaking LLMs Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 16
- Resolution
- metadata mismatch
- arxiv_id, observed 2026-05-13T07:07:27.049638Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation ca6d5dbd-88cc-414b-a563-b525b536752d · inbound
Early Data Exposure Improves Robustness to Subsequent Fine-Tuning cites this paper.
Early Data Exposure Improves Robustness to Subsequent Fine-Tuning Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 22
- Resolution
- verified exact
- arxiv_id, observed 2026-05-14T20:47:58.535416Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation 363a98b2-587f-40db-bea0-4c105b6ccaa9 · 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 Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 45
- Resolution
- metadata mismatch
- arxiv_id, observed 2026-06-30T21:05:04.065497Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation f2e55edf-6079-4b2a-bbfc-0111bee5acf3 · inbound
From AI-Generated Content to Agentic Action: Security and Safety Threats in Generative AI cites this paper.
From AI-Generated Content to Agentic Action: Security and Safety Threats in Generative AI Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 145
- Resolution
- metadata mismatch
- arxiv_id, observed 2026-05-20T18:08:50.464701Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation 0b84fdc6-85e8-4316-85e1-129d7639b7ab · inbound
Adversarial Reframing: A Framework for Targeted Generation in Language Models cites this paper.
Adversarial Reframing: A Framework for Targeted Generation in Language Models Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 57
- Resolution
- metadata mismatch
- arxiv_id, observed 2026-05-22T09:36:21.003990Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation edb54687-7a19-4cea-a398-f22302efe7ad · inbound
Palette: A Modular, Controllable, and Efficient Framework for On-demand Authorized Safety Alignment Relaxation in LLMs cites this paper.
Palette: A Modular, Controllable, and Efficient Framework for On-demand Authorized Safety Alignment Relaxation in LLMs Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 6
- Resolution
- metadata mismatch
- arxiv_id, observed 2026-06-30T16:04:52.548154Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation 7cacd102-5732-4aed-8bc9-feaef39b7a1c · inbound
Open-Weight LLM Fine-Tuning Defenses are Susceptible to Simple Attacks cites this paper.
Open-Weight LLM Fine-Tuning Defenses are Susceptible to Simple Attacks Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 17
- Resolution
- verified exact
- arxiv_id, observed 2026-06-29T19:23:53.646667Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation 42583a00-abd0-4214-a5eb-7b6dfc4c5c72 · inbound
SPARD: Defending Harmful Fine-Tuning Attack via Safety Projection with Relevance-Diversity Data Selection cites this paper.
SPARD: Defending Harmful Fine-Tuning Attack via Safety Projection with Relevance-Diversity Data Selection Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 17
- Resolution
- verified exact
- arxiv_id, observed 2026-06-29T13:53:28.660121Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation 86a45327-cef7-4f8f-bcae-13b1650f33df · inbound
Feature Geometry of LoRA Adapters: A Sparse Autoencoder Analysis of Representational Divergence in Fine-Tuned Language Models cites this paper.
Feature Geometry of LoRA Adapters: A Sparse Autoencoder Analysis of Representational Divergence in Fine-Tuned Language Models Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 16
- Resolution
- verified exact
- arxiv_id, observed 2026-06-29T13:53:28.777900Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation d0ffbb66-00a2-4a1e-8735-f2177e4e3f6b · inbound
Aligned but Fragile: Enhancing LLM Safety Robustness via Zeroth-Order Optimization cites this paper.
Aligned but Fragile: Enhancing LLM Safety Robustness via Zeroth-Order Optimization Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 36
- Resolution
- metadata mismatch
- arxiv_id, observed 2026-06-29T07:43:13.494789Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation bf2da3a5-6d73-4855-ad5b-43ac28e24aad · inbound
CSULoRA: Closest Safe Update Low-Rank Adaptation cites this paper.
CSULoRA: Closest Safe Update Low-Rank Adaptation Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 2
- Resolution
- verified exact
- arxiv_id, observed 2026-06-29T08:33:15.752828Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation 916ab45f-0e62-42bf-937a-b2dac39d344a · inbound
DataShield: Safety-degrading Data Filtering for LLM Benign Instruction Fine-Tuning cites this paper.
DataShield: Safety-degrading Data Filtering for LLM Benign Instruction Fine-Tuning Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 3
- Resolution
- metadata mismatch
- arxiv_id, observed 2026-07-01T19:46:10.220470Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation c1a7d728-736d-4eb8-b6dc-091baa690fe3 · inbound
CANARY: Zero-Label Detection of Fine-Tuning Contamination in Language Models cites this paper.
CANARY: Zero-Label Detection of Fine-Tuning Contamination in Language Models Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 17
- Resolution
- metadata mismatch
- arxiv_id, observed 2026-07-01T22:26:18.023666Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation 0d5f89a3-ee3c-4a53-9222-020f6c0cd61a · inbound
Jailbreaking Multimodal Large Language Models using Multi-Clip Video cites this paper.
Jailbreaking Multimodal Large Language Models using Multi-Clip Video Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 5
- Resolution
- verified exact
- arxiv_id, observed 2026-07-01T22:36:17.122384Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation f7687dd7-8ec9-48f2-ad93-d68d17261cf4 · inbound
Trait-space Monitoring for Emergent Misalignment During Supervised Finetuning cites this paper.
Trait-space Monitoring for Emergent Misalignment During Supervised Finetuning Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 4
- Resolution
- metadata mismatch
- arxiv_id, observed 2026-07-01T20:56:13.889072Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation d5d76cf4-66d7-428d-b646-4fc822d144a0 · inbound
Sch\"utzen: Evaluating LLM Safety in Bulgarian and German Contexts cites this paper.
Sch\"utzen: Evaluating LLM Safety in Bulgarian and German Contexts Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 132
- Resolution
- metadata mismatch
- arxiv_id, observed 2026-07-03T04:57:38.392850Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation 6b09bb8b-3d4c-4d57-b297-7c818bb3df68 · inbound
ALIGNBEAM : Inference-Time Alignment Transfer via Cross-Vocabulary Logit Mixing cites this paper.
ALIGNBEAM : Inference-Time Alignment Transfer via Cross-Vocabulary Logit Mixing Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 11
- Resolution
- metadata mismatch
- arxiv_id, observed 2026-07-03T10:27:56.364793Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation 75df2934-0219-4bd2-aefc-b5a00ad87e95 · inbound
Do Activation Monitors Survive Model Updates? Benchmarking, Predicting, and Repairing Activation-Monitor Staleness cites this paper.
Do Activation Monitors Survive Model Updates? Benchmarking, Predicting, and Repairing Activation-Monitor Staleness Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 104
- Resolution
- verified exact
- arxiv_id, observed 2026-07-03T17:58:47.686325Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation 0be25dfd-52ef-48ac-a9fa-a35c7f8ac810 · inbound
Beyond Safe Data: Pretraining-Stage Alignment with Regular Safety Reflection cites this paper.
Beyond Safe Data: Pretraining-Stage Alignment with Regular Safety Reflection Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 68
- Resolution
- metadata mismatch
- arxiv_id, observed 2026-07-04T01:09:18.893870Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation ff6e584a-c301-4845-b460-02cd23924670 · inbound
Skin-Deep: A Geometric Diagnostic for Alignment Fragility in Large Language Model Representations cites this paper.
Skin-Deep: A Geometric Diagnostic for Alignment Fragility in Large Language Model Representations Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 31
- Resolution
- metadata mismatch
- arxiv_id, observed 2026-07-04T09:09:43.661267Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation 376d516e-062b-4eae-b132-ebc3db2dad25 · inbound
Defending Against Harmful Supervision Hidden in Benign Samples cites this paper.
Defending Against Harmful Supervision Hidden in Benign Samples Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 6
- Resolution
- metadata mismatch
- arxiv_id, observed 2026-06-30T14:24:45.321547Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation 45e0734a-62f9-480b-9689-95425a50aea7 · inbound
Evil Spectra: How Optimisers can Amplify or Suppress Emergent Misalignment cites this paper.
Evil Spectra: How Optimisers can Amplify or Suppress Emergent Misalignment Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 19
- Resolution
- metadata mismatch
- arxiv_id, observed 2026-07-01T09:45:39.623496Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.
Observation a23f879c-557c-4abf-a2e4-8c4812be1ddb · inbound
Not All Refusals Are Equal: How Safety Alignment Fails Cybersecurity at Scale cites this paper.
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 6271f3ad-1a30-4338-bb64-853b857090af · inbound
TRACE: Trajectory-Based Safety Patch Learning for LLM Post-Training Realignment cites this paper.
TRACE: Trajectory-Based Safety Patch Learning for LLM Post-Training Realignment Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 15
- Resolution
- unresolved
- no resolver link, observed 2026-08-02T10:00:10.311613Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 6dc0aa6b-640a-4bd9-8fbc-554c13c04373 · inbound
How LLM Task-Adaptation Reshapes Alignment: A Multi-dimensional Study of Behavioral and Representational Drift cites this paper.
How LLM Task-Adaptation Reshapes Alignment: A Multi-dimensional Study of Behavioral and Representational Drift Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 40
- Resolution
- unresolved
- no resolver link, observed 2026-08-02T07:44:12.284016Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 933e9d9a-13e5-49af-ab4a-1157bbacc3ec · inbound
Agent Safety Should Be a Runtime Contract cites this paper.
Agent Safety Should Be a Runtime Contract Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
Reference 97
- Resolution
- unresolved
- no resolver link, observed 2026-08-15T14:19:13.984292Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.