Paper Citation Record · LEDGER

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models

As of 8 August 2026, this Paper Citation Record lists 38 of 38 outbound references and 2 inbound Pith citation observations for arXiv:2602.02600.

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

pith.paper-citation-record.v1
2602.02600 v3

Coverage vector

measured 38 of 38 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-08-03T05:45:45.089748Z

measured 40 of 40 standing notices

One-hop event checks from named stored sources.

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

measured 2 of 2 inbound itemization

Pith citing papers itemized under the disclosed page cap.

Source: paper_references, paper_reference_links, observed 2026-05-20T11:26:55.810822Z

measured 1 of 1 external citation measurements

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

Source: arxiv_reference, observed 2026-08-05T02:28:24.338817Z

Reference resolution

38 of 38 outbound references displayed

verified exact0
verified fuzzy0
unresolved37
parse uncertain0
malformed identifier1
metadata mismatch0

External citation measurements

0
arxiv_reference, observed 2026-08-05T02:28:24.338817Z

Outbound references

Observation 0af9e6f5-b904-439c-927d-71c7ec52a5a9 · outbound

This paper cites GPT-4 Technical Report.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models GPT-4 Technical Report

Reference 1

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:41.616691Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f79f4f1f-4dd5-43bb-b67b-d8c4f1cf3c08 · outbound

This paper cites Jailbreaking Leading Safety-Aligned LLMs with Simple Adaptive Attacks.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models Jailbreaking Leading Safety-Aligned LLMs with Simple Adaptive Attacks

Reference 3

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:41.783523Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation cef9de61-027f-4f8d-8796-dad0bd30655f · outbound

This paper cites How Contextual are Contextualized Word Representations? Comparing the Geometry of BERT, ELMo, and GPT-2 Embeddings.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models How Contextual are Contextualized Word Representations? Comparing the Geometry of BERT, ELMo, and GPT-2 Embeddings

Reference 6

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:41.989458Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8bde1605-6053-4eaf-a72f-0bfe42687318 · outbound

This paper cites Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations

Reference 7

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:42.088173Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 645a920c-ee84-45a8-8165-a35bf813f178 · outbound

This paper cites A2d: Any-order, any-step safety align- ment for diffusion language models.arXiv preprint arXiv:2509.23286,.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models A2d: Any-order, any-step safety align- ment for diffusion language models.arXiv preprint arXiv:2509.23286,

Reference 8

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:42.151015Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d91a8e40-f89f-4d19-b118-a621952607ab · outbound

This paper cites Crafting papers on machine learning.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models Crafting papers on machine learning

Reference 9

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:42.207501Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4aa0717e-6e43-4b42-b623-f08cf4a9df94 · outbound

This paper cites Refusal Suppression.Refusal Suppression is implemented following the prompt-based method introduced in (Wei et al., 2023).

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models Refusal Suppression.Refusal Suppression is implemented following the prompt-based method introduced in (Wei et al., 2023)

Reference 10

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:44.663357Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c45463bd-11a8-4de5-9d58-d0d0eef889d6 · outbound

This paper cites Jailbreak Attack Initializations as Extractors of Compliance Directions.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models Jailbreak Attack Initializations as Extractors of Compliance Directions

Reference 11

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:42.429452Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2ef57400-6880-4f34-bb57-99551e0497a3 · outbound

This paper cites FlipAttack: Jailbreak LLMs via Flipping.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models FlipAttack: Jailbreak LLMs via Flipping

Reference 13

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:42.577745Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 0f9c8323-108f-4d09-802f-9e189777ab62 · outbound

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

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal

Reference 14

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:42.702775Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 79833021-1f3f-4cad-8c93-f07497e42aa3 · outbound

This paper cites Large Language Diffusion Models.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models Large Language Diffusion Models

Reference 15

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:42.780302Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8270372a-f80b-497a-b771-f6f76a872fa9 · outbound

This paper cites LLM Self Defense: By Self Examination, LLMs Know They Are Being Tricked.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models LLM Self Defense: By Self Examination, LLMs Know They Are Being Tricked

Reference 16

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:42.882537Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f5353b55-d912-427a-a9e8-f0f2abcc47dd · outbound

This paper cites Large Language Model Safety: A Holistic Survey.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models Large Language Model Safety: A Holistic Survey

Reference 18

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:43.008497Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation fd1bd27b-07a2-4db1-bf4e-b921ce2f7fe8 · outbound

This paper cites Gemma: Open Models Based on Gemini Research and Technology.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models Gemma: Open Models Based on Gemini Research and Technology

Reference 19

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:43.091156Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 38181c6c-e721-4a81-8377-6ddca49e4aac · outbound

This paper cites BERT Rediscovers the Classical NLP Pipeline.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models BERT Rediscovers the Classical NLP Pipeline

Reference 20

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:43.168094Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 659cb4a2-6c44-4335-999b-26af90ba3d57 · outbound

This paper cites The devil behind the mask: An emergent safety vulnerability of diffusion llms.arXiv preprint arXiv:2507.11097,.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models The devil behind the mask: An emergent safety vulnerability of diffusion llms.arXiv preprint arXiv:2507.11097,

Reference 21

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:43.286735Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4623871c-c93f-49f9-a08b-3449b3bd9c6b · outbound

This paper cites The geometry of refusal in large language models: Concept cones and representa- tional independence.arXiv preprint arXiv:2502.17420,.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models The geometry of refusal in large language models: Concept cones and representa- tional independence.arXiv preprint arXiv:2502.17420,

Reference 22

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:43.334338Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b3b4624a-5506-431b-8382-3254edd8199a · outbound

This paper cites A Survey of Attacks on Large Language Models.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models A Survey of Attacks on Large Language Models

Reference 23

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:43.424556Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4ec446d6-0ff9-4998-95de-1f29e5e44e6d · outbound

This paper cites Qwen2.5-1M Technical Report.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models Qwen2.5-1M Technical Report

Reference 24

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:43.533522Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation bd6f25c0-3a76-4335-aa7f-b6ce870c33cf · outbound

This paper cites Dream 7B: Diffusion Large Language Models.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models Dream 7B: Diffusion Large Language Models

Reference 25

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:43.602807Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 777b2d59-949d-4018-9c6a-eaf9c7c97545 · outbound

This paper cites Discrete diffusion in large lan- guage and multimodal models: A survey.arXiv preprint arXiv:2506.13759,.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models Discrete diffusion in large lan- guage and multimodal models: A survey.arXiv preprint arXiv:2506.13759,

Reference 26

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:43.714527Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 89253ad6-01f5-4fbe-89ac-e287371ac473 · outbound

This paper cites Jailbreaking Large Language Diffusion Models: Revealing Hidden Safety Flaws in Diffusion-Based Text Generation.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models Jailbreaking Large Language Diffusion Models: Revealing Hidden Safety Flaws in Diffusion-Based Text Generation

Reference 27

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:43.832756Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6e3c96fc-248a-4283-9bd5-e37970a29c02 · outbound

This paper cites A Survey of Large Language Models.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models A Survey of Large Language Models

Reference 28

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:43.930775Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e107358e-4f46-4aef-a539-16469f26a025 · outbound

This paper cites LLaDA 1.5: Variance-Reduced Preference Optimization for Large Language Diffusion Models.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models LLaDA 1.5: Variance-Reduced Preference Optimization for Large Language Diffusion Models

Reference 29

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:44.043197Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 080d095d-ce43-4f94-abba-3ec258137151 · outbound

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

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models Universal and Transferable Adversarial Attacks on Aligned Language Models

Reference 30

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:44.104063Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b124f1db-3f47-4ee5-a44f-0a6bd0bffd4c · outbound

This paper cites In setups with active defenses, any prompt or response blocked by a defense is also counted toward the Refusal Rate.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models In setups with active defenses, any prompt or response blocked by a defense is also counted toward the Refusal Rate

Reference 31

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:44.215298Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 99a4289c-9bf2-436d-97f3-dc8b23408f56 · outbound

This paper cites True" if it is answering or.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models True" if it is answering or

Reference 32

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:44.364746Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation bd0d541b-b0c9-4ff6-b18d-c10906c6dd90 · outbound

This paper cites Text: {response} Answer: Figure 8.Judge prompt used to evaluate whether a model response contains harmful information.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models Text: {response} Answer: Figure 8.Judge prompt used to evaluate whether a model response contains harmful information

Reference 33

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:44.522702Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 33bdee98-4d52-4f47-912f-00cb694bc98c · outbound

This paper cites We use a subset of 100 PAIR-generated jailbreak prompts provided in the repository.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models We use a subset of 100 PAIR-generated jailbreak prompts provided in the repository

Reference 35

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:44.789698Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c26b6b94-e78e-47a5-8aef-4fad5fec848d · outbound

This paper cites We use a subset of 100 Random Search jailbreak prompts for evaluation.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models We use a subset of 100 Random Search jailbreak prompts for evaluation

Reference 36

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:44.896303Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 04a391ec-90cb-412f-9616-00c55a0a6440 · outbound

This paper cites For each response, we compute the average per-token negative log-likelihood under GPT-2 and reject outputs whose perplexity exceeds a fixed threshold.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models For each response, we compute the average per-token negative log-likelihood under GPT-2 and reject outputs whose perplexity exceeds a fixed threshold

Reference 37

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:45.026784Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d57c5549-f954-430d-9660-83838cb14afc · outbound

This paper cites <|begin_of_text|> [INST] Task: Check if the following content is safe or unsafe.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models <|begin_of_text|> [INST] Task: Check if the following content is safe or unsafe

Reference 38

Resolution: malformed identifier; no resolver link, observed 2026-08-03T05:45:45.089748Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c0ba13cc-7577-4cd3-9ca2-2dfa84908394 · outbound

This paper cites Interpretation Meets Safety: A Survey on Interpretation Methods and Tools for Improving LLM Safety.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models Interpretation Meets Safety: A Survey on Interpretation Methods and Tools for Improving LLM Safety

Reference 2000

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:42.313410Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d40a9abd-ada4-4e57-acef-072fcfbb3a7a · outbound

This paper cites Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks

Reference 2019

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:42.950747Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation caaa3fb1-b72d-4031-b394-2af6d2856717 · outbound

This paper cites Diffuguard: How in- trinsic safety is lost and found in diffusion large language models.arXiv preprint arXiv:2509.24296,.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models Diffuguard: How in- trinsic safety is lost and found in diffusion large language models.arXiv preprint arXiv:2509.24296,

Reference 2022

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:42.491993Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation fe89cb34-533d-47c8-9e36-339b16167c22 · outbound

This paper cites Detecting Language Model Attacks with Perplexity.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models Detecting Language Model Attacks with Perplexity

Reference 2023

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:41.660950Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 76555a29-24b3-4b02-9b3a-4a097edaea0c · outbound

This paper cites LLaDA2.0: Scaling Up Diffusion Language Models to 100B.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models LLaDA2.0: Scaling Up Diffusion Language Models to 100B

Reference 2024

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:41.842845Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1e4310c0-1c9b-4c52-9dbf-ef67e2c47430 · outbound

This paper cites D., Dhariwal, P., Neelakantan, A., Shyam, P., Sastry, G., Askell, A., et al.

Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models D., Dhariwal, P., Neelakantan, A., Shyam, P., Sastry, G., Askell, A., et al

Reference 2025

Resolution: unresolved; no resolver link, observed 2026-08-03T05:45:41.899498Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Pith citing papers

Observation 7bf662f9-f996-4d98-88ef-0c3e7c16450f · inbound

Differences in Text Generated by Diffusion and Autoregressive Language Models cites this paper.

Differences in Text Generated by Diffusion and Autoregressive Language Models Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models

Reference 28

Resolution: metadata mismatch; arxiv_id, observed 2026-06-08T02:04:45.785996Z

Source-reported events for the cited work

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

Observation e245d122-1e7f-4868-8adf-2c7c0c89f951 · inbound

SafeDiffusion-R1: Online Reward Steering for Safe Diffusion Post-Training cites this paper.

SafeDiffusion-R1: Online Reward Steering for Safe Diffusion Post-Training Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models

Reference 133

Resolution: verified exact; arxiv_id, observed 2026-06-08T02:04:45.785996Z

Source-reported events for the cited work

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