Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs | Paper Citation Record

Observation e8321ce9-81da-42ae-8ec0-c6fb8513d7d1 · outbound

This paper cites The Challenges of Evaluating LLM Applications: An Analysis of Automated, Human, and LLM-Based Approaches.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs The Challenges of Evaluating LLM Applications: An Analysis of Automated, Human, and LLM-Based Approaches

Reference 1

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.636383Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 53f87135-8398-4774-9d33-492f27ac430e · outbound

This paper cites Chatdoctor healthcaremagic-100k,.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Chatdoctor healthcaremagic-100k,

Reference 2

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.641606Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 3c571bf7-a3ad-4bae-8ec5-7cea169d8f1e · outbound

This paper cites Prompt library, 2025.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Prompt library, 2025

Reference 3

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.649022Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 00da77a2-02af-4dfe-aac6-095bd74aac49 · outbound

This paper cites Baffle: Backdoor detection via feedback-based federated learning.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Baffle: Backdoor detection via feedback-based federated learning

Reference 4

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.652692Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e27d3ecc-1b95-4298-a0c0-1425b33f9d4d · outbound

This paper cites Foundational Challenges in Assuring Alignment and Safety of Large Language Models.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Foundational Challenges in Assuring Alignment and Safety of Large Language Models

Reference 5

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.656303Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 089c15a0-3bb5-40b8-b1a5-2db23513320c · outbound

This paper cites Refusal in language models is mediated by a single direction.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Refusal in language models is mediated by a single direction

Reference 6

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.660432Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 73ce5991-957f-4f16-9ec8-b79bc4acc0ab · outbound

This paper cites Safety-tuned LLaMAs: Lessons from im- proving the safety of large language models that fol- low instructions.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Safety-tuned LLaMAs: Lessons from im- proving the safety of large language models that fol- low instructions

Reference 7

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.663947Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d84e8e31-c6ab-450c-8705-a80c560fe3b0 · outbound

This paper cites Machine learning with adver- saries: Byzantine tolerant gradient descent.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Machine learning with adver- saries: Byzantine tolerant gradient descent

Reference 8

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.667675Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2955e32b-626c-4c1e-90fc-39a56edf226e · outbound

This paper cites Scaling Trends for Data Poisoning in LLMs.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Scaling Trends for Data Poisoning in LLMs

Reference 9

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.671119Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ab5422b9-dac6-43c4-8667-a47e0221281d · outbound

This paper cites Poisoning web-scale training datasets is practi- cal.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Poisoning web-scale training datasets is practi- cal

Reference 10

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.675001Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f5efe2ce-94d8-4139-9bb3-4c14364346c8 · outbound

This paper cites Towards fed- erated large language models: Motivations, methods, and future directions.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Towards fed- erated large language models: Motivations, methods, and future directions

Reference 11

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.678606Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation dbeadc3c-9e65-4e33-8f59-6bfb950eebbd · outbound

This paper cites Llm agents for education: Advances and applications.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Llm agents for education: Advances and applications

Reference 12

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.682209Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation cddb2dcc-c2b7-4443-80fa-a509bf88389b · outbound

This paper cites Cover and Joy A.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Cover and Joy A

Reference 13

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.686579Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 5c682212-dc14-49ff-a7dd-60fcd3d7f8a3 · outbound

This paper cites I-divergence geometry of probability distributions and minimization problems.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs I-divergence geometry of probability distributions and minimization problems

Reference 14

Resolution: malformed identifier; no resolver link, observed 2026-08-15T16:51:54.690132Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation aa635d91-7530-4258-8ad7-bef7914767f1 · outbound

This paper cites Unifying bias and unfairness in information retrieval: New challenges in the llm era.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Unifying bias and unfairness in information retrieval: New challenges in the llm era

Reference 15

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.693980Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6a37e67a-09d6-4ec9-8033-90c8d2b26c30 · outbound

This paper cites the china virus.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs the china virus

Reference 16

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.697459Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 373c2f57-8145-4672-abdf-2a2a6b0e20d8 · outbound

This paper cites Qlora: Efficient finetuning of quan- tized llms.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Qlora: Efficient finetuning of quan- tized llms

Reference 17

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.700971Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation fe5ee33a-8f28-410e-8e66-4c92f80791fb · outbound

This paper cites The Philosopher's Stone: Trojaning Plugins of Large Language Models.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs The Philosopher's Stone: Trojaning Plugins of Large Language Models

Reference 18

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.704367Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f97014a1-06d0-490d-b032-93a82af1c033 · outbound

This paper cites Fairness in graph mining: A survey.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Fairness in graph mining: A survey

Reference 19

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T16:51:55.993645Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation f9a420c6-b484-4f76-b2a5-cacca3fb78a9 · outbound

This paper cites On structural explanation of bias in graph neural networks.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs On structural explanation of bias in graph neural networks

Reference 20

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T16:51:55.982381Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation 3c0741f0-1ba7-4ae4-bb2c-645cab564884 · outbound

This paper cites Attacks, Defenses and Evaluations for LLM Conversation Safety: A Survey.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Attacks, Defenses and Evaluations for LLM Conversation Safety: A Survey

Reference 21

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.714970Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7f492a03-d229-4851-9cef-978c83f39783 · outbound

This paper cites Byzantine-resilient zero-order optimization for scalable federated fine-tuning of large language models.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Byzantine-resilient zero-order optimization for scalable federated fine-tuning of large language models

Reference 22

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T16:51:55.972531Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation 8bd8bb4f-0e10-416c-94cb-ca195c0bf54f · outbound

This paper cites Freqfed: A frequency analysis-based approach for mitigating poisoning attacks in federated learning.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Freqfed: A frequency analysis-based approach for mitigating poisoning attacks in federated learning

Reference 23

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T16:51:55.961668Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation 3c421e8f-3a9a-4685-8c7b-1a3e3bc27309 · outbound

This paper cites Should ChatGPT be Biased? Challenges and Risks of Bias in Large Language Models.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Should ChatGPT be Biased? Challenges and Risks of Bias in Large Language Models

Reference 24

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.725476Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6ae72fc7-2da3-4bf7-81ba-3ed2fc684d08 · outbound

This paper cites Attack-Resistant Federated Learning with Residual-based Reweighting.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Attack-Resistant Federated Learning with Residual-based Reweighting

Reference 25

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.729793Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b1d55f92-2772-442f-8bea-b3b52f7a1ac2 · outbound

This paper cites Mitigating Sybils in Federated Learning Poisoning.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Mitigating Sybils in Federated Learning Poisoning

Reference 26

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.733817Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 76bd42a5-40ac-4796-a67c-fb35b8270bba · outbound

This paper cites Bias and fairness in large language models: A survey.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Bias and fairness in large language models: A survey

Reference 27

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T16:51:55.951085Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation 0b1e50da-a096-4b26-8db0-f657357fdc4b · outbound

This paper cites Resume dataset, 2024.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Resume dataset, 2024

Reference 28

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T16:51:55.940176Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation 162280e2-57e6-4851-8f5f-1c2249f424ec · outbound

This paper cites Application of llm agents in recruitment: a novel frame- work for automated resume screening.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Application of llm agents in recruitment: a novel frame- work for automated resume screening

Reference 29

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T16:51:55.929304Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation 7297f6f0-a653-4231-8e10-521729342a3a · outbound

This paper cites Denial-of-Service Poisoning Attacks against Large Language Models.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Denial-of-Service Poisoning Attacks against Large Language Models

Reference 30

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.748083Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 95aadeea-71f8-4758-a3cf-75e00845711e · outbound

This paper cites Patient-clinician interac- tions and disparities in breast cancer care: the equality in breast cancer care study.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Patient-clinician interac- tions and disparities in breast cancer care: the equality in breast cancer care study

Reference 31

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T16:51:55.917954Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation b689980e-871d-4235-a552-7201f357fcc5 · outbound

This paper cites BadNets: Identifying Vulnerabilities in the Machine Learning Model Supply Chain.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs BadNets: Identifying Vulnerabilities in the Machine Learning Model Supply Chain

Reference 32

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.755219Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 3a0b5139-e84a-41e3-9293-f1bc1211fd54 · outbound

This paper cites Covert Malicious Finetuning: Challenges in Safeguarding LLM Adaptation.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Covert Malicious Finetuning: Challenges in Safeguarding LLM Adaptation

Reference 33

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.758884Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 663c0a40-1364-4af4-8c53-7593969fd05f · outbound

This paper cites Fedsecurity: A benchmark for attacks and defenses in federated learn- ing and federated llms.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Fedsecurity: A benchmark for attacks and defenses in federated learn- ing and federated llms

Reference 34

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T16:51:55.906507Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation 97184f60-8615-4e34-8647-b3e965d83592 · outbound

This paper cites Parameter-Efficient Fine-Tuning for Large Models: A Comprehensive Survey.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Parameter-Efficient Fine-Tuning for Large Models: A Comprehensive Survey

Reference 35

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.765986Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d58f0374-e578-40a6-a119-f93e7efe0e60 · outbound

This paper cites Catastrophic Forgetting in LLMs: A Comparative Analysis Across Language Tasks.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Catastrophic Forgetting in LLMs: A Comparative Analysis Across Language Tasks

Reference 36

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.769701Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d9c5a6f2-2ba2-4dac-873f-17dddc0dc167 · outbound

This paper cites Refusal Behavior in Large Language Models: A Nonlinear Perspective.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Refusal Behavior in Large Language Models: A Nonlinear Perspective

Reference 37

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.773379Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 751032fa-d108-4b00-a5a7-4bc8f7058f69 · outbound

This paper cites Lora: Low-rank adaptation of large language models.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Lora: Low-rank adaptation of large language models

Reference 38

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.777141Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 63fb482d-ab58-46c1-841b-ee9b0d7817a0 · outbound

This paper cites Catastrophic Jailbreak of Open-source LLMs via Exploiting Generation.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Catastrophic Jailbreak of Open-source LLMs via Exploiting Generation

Reference 39

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.780472Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 5b537448-f411-4c53-b5fb-1ddab1e47ce8 · outbound

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

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training

Reference 40

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.784207Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8743e35b-2354-48a5-803b-696cb6ca98ec · outbound

This paper cites Gpt-4o: The cutting-edge advancement in multimodal llm.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Gpt-4o: The cutting-edge advancement in multimodal llm

Reference 41

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.788255Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a358aa0d-65aa-4d80-8008-29cad90f8613 · outbound

This paper cites Mesas: Poi- soning defense for federated learning resilient against adaptive attackers.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Mesas: Poi- soning defense for federated learning resilient against adaptive attackers

Reference 42

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T16:51:55.881737Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation 3b05bd9d-12c8-4c1f-a818-54937104a9d9 · outbound

This paper cites A literature survey on open source large language models.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs A literature survey on open source large language models

Reference 43

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T16:51:55.870456Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation d5639a06-d99e-4775-809d-bcfd832a40ca · outbound

This paper cites SALAD-Bench: A Hierarchical and Comprehensive Safety Benchmark for Large Language Models.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs SALAD-Bench: A Hierarchical and Comprehensive Safety Benchmark for Large Language Models

Reference 44

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.798464Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 0b23d1c7-75d5-4f05-946a-cfee0f0ef679 · outbound

This paper cites Backdoorllm: A comprehensive benchmark for backdoor attacks on large language models.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Backdoorllm: A comprehensive benchmark for backdoor attacks on large language models

Reference 45

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T16:51:55.858993Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation 9eeb720b-97bd-4da9-8c0e-97b2ca4b5fee · outbound

This paper cites Chatdoctor: A medical chat model fine-tuned on a large language model meta-ai (llama) using medical domain knowledge.Cureus, 15(6), 2023.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Chatdoctor: A medical chat model fine-tuned on a large language model meta-ai (llama) using medical domain knowledge.Cureus, 15(6), 2023

Reference 46

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.805577Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d2495173-2bd3-4068-81e9-86c9aa6ee9e6 · outbound

This paper cites An Empirical Study of Catastrophic Forgetting in Large Language Models During Continual Fine-tuning.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs An Empirical Study of Catastrophic Forgetting in Large Language Models During Continual Fine-tuning

Reference 47

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.808979Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b510088c-d8cf-4c37-bb0a-2ceb555dc5dd · outbound

This paper cites Vicarious racism stress and disease activity: the black women’s experiences living with lupus (bewell) study.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Vicarious racism stress and disease activity: the black women’s experiences living with lupus (bewell) study

Reference 48

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T16:51:55.837854Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation f43e3977-08a8-4fad-9062-a4b47cb5ddf4 · outbound

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

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal

Reference 49

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.816029Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 91614384-6559-44ce-8101-c608d1e2c22d · outbound

This paper cites Communication- efficient learning of deep networks from decentralized data.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Communication- efficient learning of deep networks from decentralized data

Reference 50

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.819873Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c1b8dc7e-d3e6-4d4e-8069-a26ce3210ca2 · outbound

This paper cites Exploring us shifts in anti-asian sentiment with the emergence of covid-19.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Exploring us shifts in anti-asian sentiment with the emergence of covid-19

Reference 51

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T16:51:55.819424Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation 4e8b607f-26a5-47e7-b668-c54fd77e8f45 · outbound

This paper cites Training language models to follow instructions with human feedback.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Training language models to follow instructions with human feedback

Reference 52

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.826722Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f4afd77e-a7ed-43ec-81d3-6185cd27af19 · outbound

This paper cites Is poisoning a real threat to LLM alignment? Maybe more so than you think.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Is poisoning a real threat to LLM alignment? Maybe more so than you think

Reference 53

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.830133Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b0d6c281-3264-4c93-a7f1-15352c3cdba1 · outbound

This paper cites The RefinedWeb Dataset for Falcon LLM: Outperforming Curated Corpora with Web Data, and Web Data Only.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs The RefinedWeb Dataset for Falcon LLM: Outperforming Curated Corpora with Web Data, and Web Data Only

Reference 54

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.834023Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 95edf3d8-f934-41c7-9941-e5b5de4881e5 · outbound

This paper cites ONION: A Simple and Effective Defense Against Textual Backdoor Attacks.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs ONION: A Simple and Effective Defense Against Textual Backdoor Attacks

Reference 55

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.838503Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4693a88f-adf8-4d1b-b474-49d1eb553207 · outbound

This paper cites Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To!.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To!

Reference 56

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.842447Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 52f9c484-1553-4d06-9648-6037a876643b · outbound

This paper cites Hsf: Defending against jailbreak attacks with hidden state filtering.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Hsf: Defending against jailbreak attacks with hidden state filtering

Reference 57

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T16:51:55.800390Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation 52cc392c-f8d0-472f-8bc9-b54d72060986 · outbound

This paper cites CrowdGuard: Federated Backdoor Detection in Federated Learning.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs CrowdGuard: Federated Backdoor Detection in Federated Learning

Reference 58

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.850337Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 906bc3b9-ecdf-4e95-8861-488ba2d5766d · outbound

This paper cites DeepSight: Mitigating Backdoor Attacks in Federated Learning Through Deep Model Inspection.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs DeepSight: Mitigating Backdoor Attacks in Federated Learning Through Deep Model Inspection

Reference 59

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.854621Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 26547ed4-c29b-47af-9173-ae88af17f9b9 · outbound

This paper cites SmoothLLM: Defending Large Language Models Against Jailbreaking Attacks.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs SmoothLLM: Defending Large Language Models Against Jailbreaking Attacks

Reference 60

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.858410Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a7fc272b-56f6-42ac-a765-8ba2ed6e02b8 · outbound

This paper cites yahma/alpaca-cleaned, 2024.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs yahma/alpaca-cleaned, 2024

Reference 61

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T16:51:55.788811Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation 32069c96-ab3b-47d3-8b89-20483439d018 · outbound

This paper cites Chal- lenging fairness: A comprehensive exploration of bias in llm-based recommendations.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Chal- lenging fairness: A comprehensive exploration of bias in llm-based recommendations

Reference 62

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T16:51:55.775528Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation 8c952fdd-3e8a-4320-9e2c-ac7603876651 · outbound

This paper cites Survey of Vulnerabilities in Large Language Models Revealed by Adversarial Attacks.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Survey of Vulnerabilities in Large Language Models Revealed by Adversarial Attacks

Reference 63

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.876120Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation bd84825a-8df0-4e49-8a40-a174820a1260 · outbound

This paper cites Manipulating the byzantine: Optimizing model poisoning attacks and defenses for federated learning.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Manipulating the byzantine: Optimizing model poisoning attacks and defenses for federated learning

Reference 64

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.881084Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9e7f147b-a1e8-4914-a756-d061e5bcb780 · outbound

This paper cites Evaluating the Social Impact of Generative AI Systems in Systems and Society.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Evaluating the Social Impact of Generative AI Systems in Systems and Society

Reference 65

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.885869Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 454e06ca-8c26-47df-a70b-0a32685a74ca · outbound

This paper cites TrustLLM: Trustworthiness in Large Language Models.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs TrustLLM: Trustworthiness in Large Language Models

Reference 66

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.890597Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b14d3b43-a6b3-41e0-a4ec-6686674189b2 · outbound

This paper cites Peftguard: detecting backdoor attacks against parameter- efficient fine-tuning.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Peftguard: detecting backdoor attacks against parameter- efficient fine-tuning

Reference 67

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T16:51:55.757318Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation bc7eee9b-0f92-4fbd-9818-36ca039ba153 · outbound

This paper cites Stanford alpaca: An instruction- following llama model, 2023.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Stanford alpaca: An instruction- following llama model, 2023

Reference 68

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T16:51:55.745803Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation 2a29fb23-926c-4bfa-8ce3-d2ddba24a856 · outbound

This paper cites Fairness matters: A look at llm- generated group recommendations.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Fairness matters: A look at llm- generated group recommendations

Reference 69

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T16:51:55.733943Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation db58f4e3-5201-460d-a520-dca85bcaa9f7 · outbound

This paper cites LLaMA: Open and Efficient Foundation Language Models.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs LLaMA: Open and Efficient Foundation Language Models

Reference 70

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.908193Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c0f62e13-0286-4433-a7f4-ce8f59cf74a3 · outbound

This paper cites Padbench, 2025.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Padbench, 2025

Reference 71

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T16:51:55.721610Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation a7e87650-641a-4462-aec4-216813ecbe3a · outbound

This paper cites Poisoning language models during instruction tuning.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Poisoning language models during instruction tuning

Reference 72

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T16:51:55.709384Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation 2c0dd3d3-8dc2-41cc-a1d4-ee83ae1ccf91 · outbound

This paper cites Hybrid Alignment Training for Large Language Models.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Hybrid Alignment Training for Large Language Models

Reference 73

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.921952Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4f8201c1-613d-4bc6-9699-fd12e0cf2845 · outbound

This paper cites Backdooralign: Mitigat- ing fine-tuning based jailbreak attack with backdoor enhanced safety alignment.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Backdooralign: Mitigat- ing fine-tuning based jailbreak attack with backdoor enhanced safety alignment

Reference 74

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T16:51:55.696951Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation 1fc85b38-acb0-4778-9199-f5ea395dd297 · outbound

This paper cites Foundation Models in Federated Learning: Assessing Backdoor Vulnerabilities.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Foundation Models in Federated Learning: Assessing Backdoor Vulnerabilities

Reference 75

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.931467Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 0679a9da-40ac-4b87-8774-a89f8094c149 · outbound

This paper cites Detecting back- door attacks in federated learning via direction align- ment inspection.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Detecting back- door attacks in federated learning via direction align- ment inspection

Reference 76

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T16:51:55.683038Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation 580d3c19-fd49-4aa3-a967-1f4d9ebef8e2 · outbound

This paper cites Backdooring Instruction-Tuned Large Language Models with Virtual Prompt Injection.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Backdooring Instruction-Tuned Large Language Models with Virtual Prompt Injection

Reference 77

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.940823Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9176602f-6e5d-41d2-96e8-7f4bd45a44f9 · outbound

This paper cites Emerging safety attack and defense in federated instruction tuning of large lan- guage models.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Emerging safety attack and defense in federated instruction tuning of large lan- guage models

Reference 78

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T16:51:55.670837Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation 4f81b681-ec55-4df4-8846-35eb76f4d1f6 · outbound

This paper cites Understanding Refusal in Language Models with Sparse Autoencoders.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Understanding Refusal in Language Models with Sparse Autoencoders

Reference 79

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.950442Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1f63d085-240e-4322-9aff-635d01fe03ae · outbound

This paper cites Badacts: A universal backdoor de- fense in the activation space.Findings of the Association for Computational Linguistics: ACL 2024, 2024.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Badacts: A universal backdoor de- fense in the activation space.Findings of the Association for Computational Linguistics: ACL 2024, 2024

Reference 80

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T16:51:55.659439Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation 15e21656-9e67-4563-a7cc-8233a5166ccf · outbound

This paper cites Byzantine-robust distributed learning: Towards optimal statistical rates.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Byzantine-robust distributed learning: Towards optimal statistical rates

Reference 81

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T16:51:55.647006Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation 50b8a017-09a9-476e-82fa-2173ed6903bb · outbound

This paper cites CLIBE: Detecting dynamic back- doors in transformer-based nlp models.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs CLIBE: Detecting dynamic back- doors in transformer-based nlp models

Reference 82

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T16:51:55.635556Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation 2a43c7fe-abc1-4ec4-9863-4759915f00b2 · outbound

This paper cites Persistent Pre-Training Poisoning of LLMs.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Persistent Pre-Training Poisoning of LLMs

Reference 83

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.969388Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d12a7817-57a1-460c-95f7-ee4881b46508 · outbound

This paper cites Learning and Forgetting Unsafe Examples in Large Language Models.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Learning and Forgetting Unsafe Examples in Large Language Models

Reference 84

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.975579Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e6332d89-bc9f-40d1-a1d2-81d7bceab443 · outbound

This paper cites GPTBIAS: A Comprehensive Framework for Evaluating Bias in Large Language Models.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs GPTBIAS: A Comprehensive Framework for Evaluating Bias in Large Language Models

Reference 85

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.980320Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7d8afa5e-51ff-4216-8c24-fdfbd2b55bb5 · outbound

This paper cites Judging llm-as- a-judge with mt-bench and chatbot arena.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Judging llm-as- a-judge with mt-bench and chatbot arena

Reference 86

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.984985Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d78fd001-b955-43a3-9083-e9243b71e0c4 · outbound

This paper cites Exposing the Ghost in the Transformer: Abnormal Detection for Large Language Models via Hidden State Forensics.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Exposing the Ghost in the Transformer: Abnormal Detection for Large Language Models via Hidden State Forensics

Reference 87

Resolution: verified exact; local_arxiv, observed 2026-08-15T16:51:55.094730Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation 85669597-3741-4901-888b-f3b353235182 · outbound

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

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Universal and Transferable Adversarial Attacks on Aligned Language Models

Reference 88

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.996034Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b53c36f2-603f-46f8-bba2-4bfa96edf11f · outbound

This paper cites Consider min π: π(Rx|x)=α KL π(·| x)∥ π0(·| x).

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Consider min π: π(Rx|x)=α KL π(·| x)∥ π0(·| x)

Reference 90

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T16:51:55.617215Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation ae4cab25-0222-424f-adb1-9285cd405d1f · outbound

This paper cites an unresolved cited work.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Unresolved cited work

Reference 91

Resolution: unresolved; raw_fallback, observed 2026-08-15T16:51:55.606062Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation c49ddc30-0d03-4365-8eb9-2f099f40a4cb · outbound

This paper cites increase.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs increase

Reference 92

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T16:51:55.593148Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

Observation ddf6c29e-9790-4ef1-ae19-43c4beb22c35 · outbound

This paper cites an unresolved cited work.

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs Unresolved cited work

Reference 2024

Resolution: unresolved; no resolver link, observed 2026-08-15T16:51:54.645194Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Coverage vector

Reference resolution

Outbound references

Pith citing papers