Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions | Paper Citation Record

Observation b90455ae-f6db-4334-8b9b-4dff47035811 · outbound

This paper cites GPT-4 Technical Report.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions GPT-4 Technical Report

Reference 1

Resolution: unresolved; no resolver link, observed 2026-08-08T19:07:35.351848Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b33bec3d-afa7-4659-a17d-3d8fe19b393a · outbound

This paper cites DeepSeek-V3 Technical Report.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions DeepSeek-V3 Technical Report

Reference 2

Resolution: unresolved; no resolver link, observed 2026-08-08T19:07:35.355831Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 44078d6b-9dfb-4df5-88da-a17a08d7b3e1 · outbound

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

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions LLaMA: Open and Efficient Foundation Language Models

Reference 3

Resolution: unresolved; no resolver link, observed 2026-08-08T19:07:35.359608Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f07638cb-ec24-41fe-80bb-cdcd8e3a2928 · outbound

This paper cites Constitutional AI: Harmlessness from AI Feedback.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Constitutional AI: Harmlessness from AI Feedback

Reference 4

Resolution: unresolved; no resolver link, observed 2026-08-08T19:07:35.363738Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 34533a29-e8b8-41ee-9f1e-e620a028d73b · outbound

This paper cites The ai alignment problem: why it is hard, and where to start,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions The ai alignment problem: why it is hard, and where to start,

Reference 5

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:36.121774Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 0cc7659b-d836-47d5-9b15-275ffbaebbe6 · outbound

This paper cites Artificial intelligence, values, and alignment,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Artificial intelligence, values, and alignment,

Reference 6

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:36.111997Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 1c90bfd9-124d-4139-9ae0-dca01410a620 · outbound

This paper cites Self-instruct: Aligning language models with self- generated instructions,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Self-instruct: Aligning language models with self- generated instructions,

Reference 7

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:36.101679Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 27b1f92f-53ec-4980-af67-70b609d027eb · outbound

This paper cites Self-alignment of large language models via monopolylogue-based social scene simulation,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Self-alignment of large language models via monopolylogue-based social scene simulation,

Reference 8

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:36.091151Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 648bc372-db72-460b-ba3d-ef72a90a9c22 · outbound

This paper cites Mixture of insighTful Experts (MoTE): The Synergy of Thought Chains and Expert Mixtures in Self-Alignment.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Mixture of insighTful Experts (MoTE): The Synergy of Thought Chains and Expert Mixtures in Self-Alignment

Reference 9

Resolution: unresolved; no resolver link, observed 2026-08-08T19:07:35.381498Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d1901f4e-4d05-43f8-9993-53917d1a0576 · outbound

This paper cites Balancing differential privacy and utility: A relevance-based adaptive private fine-tuning framework for language models,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Balancing differential privacy and utility: A relevance-based adaptive private fine-tuning framework for language models,

Reference 10

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:36.082574Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation efcc71b5-2986-463c-80ce-ffc8f2c53bc7 · outbound

This paper cites Hard adversarial example mining for improving robust fairness,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Hard adversarial example mining for improving robust fairness,

Reference 11

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:36.073695Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 76e5cd17-b69e-40a1-b2d1-2e1afcbc0404 · outbound

This paper cites Finetuned language models are zero-shot learners,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Finetuned language models are zero-shot learners,

Reference 12

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:36.064386Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 805dd60b-66fe-4299-86d3-15c85574f5ed · outbound

This paper cites Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback

Reference 13

Resolution: unresolved; no resolver link, observed 2026-08-08T19:07:35.395383Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 5bde28a1-a72a-4264-bcc1-2f435ef56ae6 · outbound

This paper cites Fine-tuning aligned language models compromises safety, even when users do not intend to!.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Fine-tuning aligned language models compromises safety, even when users do not intend to!

Reference 14

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:36.054791Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 39e23c0a-ca4b-435c-9abf-ce5dfc04f8f5 · outbound

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

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Stanford alpaca: An instruction-following llama model,

Reference 15

Resolution: unresolved; no resolver link, observed 2026-08-08T19:07:35.403391Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 34156006-5528-4174-af5e-f2ab8bc7a871 · outbound

This paper cites Vicuna: An open-source chatbot impressing gpt-4 with 90%* chatgpt quality,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Vicuna: An open-source chatbot impressing gpt-4 with 90%* chatgpt quality,

Reference 16

Resolution: unresolved; no resolver link, observed 2026-08-08T19:07:35.406934Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 00152d79-6299-4ce9-a8f6-718417e69234 · outbound

This paper cites Self-instruct: Aligning language models with self- generated instructions,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Self-instruct: Aligning language models with self- generated instructions,

Reference 17

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:36.033196Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation e2c4b06d-7662-4578-8b26-e836036081b2 · outbound

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

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Training language models to follow instructions with human feedback,

Reference 18

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:36.023714Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 8399b0d9-c134-414d-bce5-e7dddf93432e · outbound

This paper cites Beavertails: Towards improved safety alignment of llm via a human-preference dataset,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Beavertails: Towards improved safety alignment of llm via a human-preference dataset,

Reference 19

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:36.014430Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 1345c631-0cc0-4ecf-be07-6b52b9a6b7f5 · outbound

This paper cites Poisoning language models during instruction tuning,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Poisoning language models during instruction tuning,

Reference 20

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:36.005000Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation fdbce315-085b-4fea-be85-68c1556a2737 · outbound

This paper cites Openassistant conversations-democratizing large language model align- ment,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Openassistant conversations-democratizing large language model align- ment,

Reference 21

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:35.994753Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 4b048045-507b-4655-9e95-31ed27700854 · outbound

This paper cites Principle-driven self-alignment of language models from scratch with minimal human supervision,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Principle-driven self-alignment of language models from scratch with minimal human supervision,

Reference 22

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:35.985261Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation d1dbda63-7d4d-4e17-a183-ab31d615a3df · outbound

This paper cites RAIN: Your Language Models Can Align Themselves without Finetuning.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions RAIN: Your Language Models Can Align Themselves without Finetuning

Reference 23

Resolution: unresolved; no resolver link, observed 2026-08-08T19:07:35.430899Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 0da6d42f-2839-4d31-8723-13b76ff2c4d6 · outbound

This paper cites RLAIF vs. RLHF: Scaling Reinforcement Learning from Human Feedback with AI Feedback.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions RLAIF vs. RLHF: Scaling Reinforcement Learning from Human Feedback with AI Feedback

Reference 24

Resolution: unresolved; no resolver link, observed 2026-08-08T19:07:35.434529Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 721bf31d-5100-4e67-830a-f891a999e833 · outbound

This paper cites Self-concept clarity development across the lifespan,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Self-concept clarity development across the lifespan,

Reference 25

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:35.975277Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 8cec7ea0-ce83-4d76-89ee-807450d21061 · outbound

This paper cites Spread Preference Annotation: Direct Preference Judgment for Efficient LLM Alignment.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Spread Preference Annotation: Direct Preference Judgment for Efficient LLM Alignment

Reference 26

Resolution: unresolved; no resolver link, observed 2026-08-08T19:07:35.441458Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 67d604db-332f-4d2e-a985-bce45b67d65c · outbound

This paper cites Secrets of RLHF in Large Language Models Part II: Reward Modeling.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Secrets of RLHF in Large Language Models Part II: Reward Modeling

Reference 27

Resolution: unresolved; no resolver link, observed 2026-08-08T19:07:35.445137Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8a8c2ee2-4adb-404e-af5b-88f0a39a02f2 · outbound

This paper cites Variational bayesian un- learning,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Variational bayesian un- learning,

Reference 28

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:35.965834Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation f09253eb-a71a-4520-8631-7bcd66c226f4 · outbound

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

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions An Empirical Study of Catastrophic Forgetting in Large Language Models During Continual Fine-tuning

Reference 29

Resolution: unresolved; no resolver link, observed 2026-08-08T19:07:35.450781Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4f879638-9652-4351-baca-4f6a1d159684 · outbound

This paper cites A Survey of Large Language Models.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions A Survey of Large Language Models

Reference 30

Resolution: unresolved; no resolver link, observed 2026-08-08T19:07:35.453762Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 92cdd58e-ad1c-4c04-8397-27b7704acb45 · outbound

This paper cites On the Opportunities and Risks of Foundation Models.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions On the Opportunities and Risks of Foundation Models

Reference 31

Resolution: unresolved; no resolver link, observed 2026-08-08T19:07:35.456785Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 55eb9564-193f-4d61-aa02-4d7628961252 · outbound

This paper cites Safety-tuned llamas: Lessons from improving the safety of large language models that follow instructions,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Safety-tuned llamas: Lessons from improving the safety of large language models that follow instructions,

Reference 32

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:35.955845Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation f9b95246-da35-47f7-a07a-4b72592d8596 · outbound

This paper cites Fine-grained human feedback gives better rewards for language model training,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Fine-grained human feedback gives better rewards for language model training,

Reference 33

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:35.944549Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation c5f597f0-13b9-422e-87bb-785a3a9c9b4e · outbound

This paper cites Learning to summarize with human feedback,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Learning to summarize with human feedback,

Reference 34

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:35.933971Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 57659805-3e28-4c17-a855-6d5f70f22100 · outbound

This paper cites Rl4f: Generating natural language feedback with reinforcement learning for repairing model outputs,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Rl4f: Generating natural language feedback with reinforcement learning for repairing model outputs,

Reference 35

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:35.925355Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 77cd4e56-c207-465f-b4b4-0d46f6b51d63 · outbound

This paper cites Gaining wisdom from setbacks: Aligning large language models via mistake analysis,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Gaining wisdom from setbacks: Aligning large language models via mistake analysis,

Reference 36

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:35.916201Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 91041920-d5a2-45a2-8069-8d8b760dae0f · outbound

This paper cites Understanding negative samples in instance dis- criminative self-supervised representation learning,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Understanding negative samples in instance dis- criminative self-supervised representation learning,

Reference 37

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:35.907055Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation c4f2e0d2-7b2f-4c36-a474-aa7acc360373 · outbound

This paper cites Learning From Failure: Integrating Negative Examples when Fine-tuning Large Language Models as Agents.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Learning From Failure: Integrating Negative Examples when Fine-tuning Large Language Models as Agents

Reference 38

Resolution: unresolved; no resolver link, observed 2026-08-08T19:07:35.479808Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 00f9962e-41d9-4b82-8d17-eebc325c02ee · outbound

This paper cites What makes for good views for contrastive learning?.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions What makes for good views for contrastive learning?

Reference 39

Resolution: unresolved; no resolver link, observed 2026-08-08T19:07:35.483473Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6ea4fda9-0edd-4acf-b225-23184ce5552e · outbound

This paper cites Neural text degeneration with unlikelihood training,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Neural text degeneration with unlikelihood training,

Reference 40

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:35.891016Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 4cde7402-0787-42a0-8e7b-cf85aa83ef18 · outbound

This paper cites Openai. gpt-4v(ision) system card,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Openai. gpt-4v(ision) system card,

Reference 41

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:35.881487Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation b78a06e6-75c4-4d4a-9fb6-f9ec1518778e · outbound

This paper cites Llama 2: Open Foundation and Fine-Tuned Chat Models.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Llama 2: Open Foundation and Fine-Tuned Chat Models

Reference 42

Resolution: unresolved; no resolver link, observed 2026-08-08T19:07:35.492858Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2c835252-f39b-495d-b08e-1a973428ef69 · outbound

This paper cites Red teaming language models to reduce harms: Methods, scaling behaviors, and lessons learned,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Red teaming language models to reduce harms: Methods, scaling behaviors, and lessons learned,

Reference 43

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:35.871580Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation a4bbf00f-7f71-48e1-8023-4e5969b47e7e · outbound

This paper cites Gemini: A Family of Highly Capable Multimodal Models.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Gemini: A Family of Highly Capable Multimodal Models

Reference 44

Resolution: unresolved; no resolver link, observed 2026-08-08T19:07:35.499871Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b9118a71-173d-4d90-94af-3407ca079fd2 · outbound

This paper cites Chain-of-thought prompting elicits reasoning in large language models,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Chain-of-thought prompting elicits reasoning in large language models,

Reference 45

Resolution: unresolved; no resolver link, observed 2026-08-08T19:07:35.503523Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 23e369c3-388c-4625-afc6-ab1058cce931 · outbound

This paper cites The wisdom of hindsight makes language models better instruction followers,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions The wisdom of hindsight makes language models better instruction followers,

Reference 46

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:35.855924Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 340786a3-7012-4239-9582-dabbf100e490 · outbound

This paper cites Koala: A dialogue model for academic research,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Koala: A dialogue model for academic research,

Reference 47

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:35.846317Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 61f8cb44-e1ea-464a-900f-ba2eac9be164 · outbound

This paper cites Glm-130b: An open bilingual pre-trained model,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Glm-130b: An open bilingual pre-trained model,

Reference 48

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:35.836713Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation a4e281db-3fdd-4ff6-ada7-49ab27538497 · outbound

This paper cites The curious case of neural text degeneration,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions The curious case of neural text degeneration,

Reference 49

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:35.826185Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 0e9c7d08-786e-4f8e-9bac-9a6946ca5f81 · outbound

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

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Lora: Low-rank adaptation of large language models,

Reference 50

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:35.815745Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 9ef50156-efc3-498c-a623-37771202d5ad · outbound

This paper cites The Llama 3 Herd of Models.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions The Llama 3 Herd of Models

Reference 51

Resolution: unresolved; no resolver link, observed 2026-08-08T19:07:35.523090Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 3fdce5e9-8d92-4dc6-9d1d-0bf8844b6519 · outbound

This paper cites Beyond the imitation game: Quantifying and extrapolating the capabilities of language models,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Beyond the imitation game: Quantifying and extrapolating the capabilities of language models,

Reference 52

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:35.805988Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation dd711585-734e-43ee-8ba3-4e08b50fd056 · outbound

This paper cites Autodan: Generating stealthy jailbreak prompts on aligned large language models,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Autodan: Generating stealthy jailbreak prompts on aligned large language models,

Reference 53

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:35.796019Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation cf6c2c09-2ba9-4856-9d5f-4b273dcf9166 · outbound

This paper cites Harmbench: A standardized evaluation framework for automated red teaming and robust refusal,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Harmbench: A standardized evaluation framework for automated red teaming and robust refusal,

Reference 54

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:35.786305Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation cf95be37-1915-4c0a-8207-3ac6d74ec4e3 · outbound

This paper cites Truthfulqa: Measuring how models mimic human falsehoods,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Truthfulqa: Measuring how models mimic human falsehoods,

Reference 55

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:35.776407Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 937b1e5a-7f3e-4358-b548-0f81431ab153 · outbound

This paper cites Measuring massive multitask language understanding,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Measuring massive multitask language understanding,

Reference 56

Resolution: unresolved; no resolver link, observed 2026-08-08T19:07:35.539940Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b23b4cd2-24f3-4aec-a5ae-6609841d685b · outbound

This paper cites Social iqa: Commonsense reasoning about social interactions,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Social iqa: Commonsense reasoning about social interactions,

Reference 57

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:35.761578Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 5534bcea-024b-408b-bde5-a85d0ea447b1 · outbound

This paper cites Mitigating the alignment tax of rlhf,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Mitigating the alignment tax of rlhf,

Reference 58

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T19:07:35.750924Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 712610c3-5c13-4414-9edf-2fd41067a154 · outbound

This paper cites Direct preference optimization: Your language model is secretly a reward model,.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions Direct preference optimization: Your language model is secretly a reward model,

Reference 59

Resolution: unresolved; no resolver link, observed 2026-08-08T19:07:35.550161Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1bfb14e8-30c3-4297-bb0d-779945f7f404 · outbound

This paper cites KTO: Model Alignment as Prospect Theoretic Optimization.

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions KTO: Model Alignment as Prospect Theoretic Optimization

Reference 60

Resolution: unresolved; no resolver link, observed 2026-08-08T19:07:35.553049Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Coverage vector

Reference resolution

Outbound references

Pith citing papers