Paper Citation Record · LEDGER

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs

As of 14 August 2026, this Paper Citation Record lists 39 of 39 outbound references and 0 inbound Pith citation observations for arXiv:2608.08542.

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

pith.paper-citation-record.v1
2608.08542 v1

Coverage vector

measured 39 of 39 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-08-14T04:46:16.031718Z

measured 39 of 39 standing notices

One-hop event checks from named stored sources.

Source: scholarly_work_events, retraction_status_cache, observed 2026-08-14T06:32:32.682623+00:00

measured 0 of 0 inbound itemization

Pith citing papers itemized under the disclosed page cap.

Source: paper_references, paper_reference_links

measured 0 of 1 external citation measurements

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

Source: cited_works

Reference resolution

39 of 39 outbound references displayed

verified exact2
verified fuzzy1
unresolved35
parse uncertain0
malformed identifier0
metadata mismatch1

External citation measurements

No source-named external measurement is stored.

Outbound references

Observation ea4a5626-c8c3-4e70-8829-b18cfd35a061 · outbound

This paper cites SafeMERGE head-to-head (main paper§7).We compare against the selective-layer baseline Safe- MERGE (Djuhera et al.

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs SafeMERGE head-to-head (main paper§7).We compare against the selective-layer baseline Safe- MERGE (Djuhera et al

Reference 1

Resolution: verified fuzzy; raw_fallback, observed 2026-08-14T04:46:16.751445Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation 5b9e2e57-c5aa-42e1-ac6f-76b9d5c56953 · outbound

This paper cites Language Models are Homer Simpson! Safety Re-Alignment of Fine-tuned Language Models through Task Arithmetic.

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs Language Models are Homer Simpson! Safety Re-Alignment of Fine-tuned Language Models through Task Arithmetic

Reference 3

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:15.901261Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 15b2f3a2-c95f-4eb2-b658-4266ccfd58fb · outbound

This paper cites JailbreakBench: An Open Robustness Benchmark for Jailbreaking Large Language Models.

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs JailbreakBench: An Open Robustness Benchmark for Jailbreaking Large Language Models

Reference 4

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:15.906278Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 48c50b2d-4fdf-4c95-b322-8adc64ac5459 · outbound

This paper cites Training Verifiers to Solve Math Word Problems.

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs Training Verifiers to Solve Math Word Problems

Reference 7

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:15.918373Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f9a16946-6a01-4984-8eba-7ea33532729f · outbound

This paper cites SafeMERGE: Preserving Safety Alignment in Fine-Tuned Large Language Models via Selective Layer-Wise Model Merging.

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs SafeMERGE: Preserving Safety Alignment in Fine-Tuned Large Language Models via Selective Layer-Wise Model Merging

Reference 9

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:15.925093Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 263505ab-b217-4b2e-b25a-28d22d546b0b · outbound

This paper cites Safeguard Fine-Tuned LLMs Through Pre- and Post-Tuning Model Merging.

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs Safeguard Fine-Tuned LLMs Through Pre- and Post-Tuning Model Merging

Reference 10

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:15.928433Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 160a6b59-3d94-4591-99d9-e8d9d0612bba · outbound

This paper cites Merging Improves Self-Critique Against Jailbreak Attacks.

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs Merging Improves Self-Critique Against Jailbreak Attacks

Reference 11

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:15.931694Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 43554abd-4fb4-4256-9cc0-980f1fe26210 · outbound

This paper cites Model Merging and Safety Alignment: One Bad Model Spoils the Bunch.

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs Model Merging and Safety Alignment: One Bad Model Spoils the Bunch

Reference 12

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:15.935356Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 5fb6fffd-c50d-49bc-81ba-9498562ad1d6 · outbound

This paper cites InAdvances in Neural Information Processing Systems (NeurIPS).

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs InAdvances in Neural Information Processing Systems (NeurIPS)

Reference 13

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:15.938898Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 10883db3-9c84-4ec1-bad8-edea276f3489 · outbound

This paper cites Measuring Massive Multitask Language Understanding.

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs Measuring Massive Multitask Language Understanding

Reference 14

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:15.941922Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8de3ed6b-2a59-4fd7-9736-82895f6c5e3e · outbound

This paper cites Safe LoRA: the Silver Lining of Reducing Safety Risks when Fine-tuning Large Language Models.

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs Safe LoRA: the Silver Lining of Reducing Safety Risks when Fine-tuning Large Language Models

Reference 15

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:15.945333Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 97bca837-2cde-42f6-8e2d-0e43a7d6a0a2 · outbound

This paper cites Editing Models with Task Arithmetic.

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs Editing Models with Task Arithmetic

Reference 16

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:15.948851Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4d59b9b3-4460-4231-8a1c-bdae26f7acce · outbound

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

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations

Reference 17

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:15.952662Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ed595a6f-298f-4b54-97b1-b072efecf5e5 · outbound

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

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs LoRA Fine-tuning Efficiently Undoes Safety Training in Llama 2-Chat 70B

Reference 18

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:15.956173Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f062de73-6dba-4e4b-b2f7-661c6350bfc7 · outbound

This paper cites Ma, Q.; Liu, D.; Chen, Q.; Zhang, L.; and Shao, J.

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs Ma, Q.; Liu, D.; Chen, Q.; Zhang, L.; and Shao, J

Reference 19

Resolution: verified exact; raw_fallback, observed 2026-08-14T04:46:16.480680Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation 34fe08dc-c22f-40d1-b7a6-97b24bf6ac3d · outbound

This paper cites LED-Merging: Mitigating Safety-Utility Conflicts in Model Merging with Location-Election-Disjoint.

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs LED-Merging: Mitigating Safety-Utility Conflicts in Model Merging with Location-Election-Disjoint

Reference 20

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:15.963181Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 86e8fea2-8046-4f65-b74d-eb2de5219de0 · outbound

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

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal

Reference 21

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:15.966829Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b5a8c56f-75eb-4375-ae08-1c6007877dbb · outbound

This paper cites Safety Alignment Should Be Made More Than Just a Few Tokens Deep.

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 22

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:15.970418Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6326a994-2fc5-488f-ba26-9a40adc8c0bb · outbound

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

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To!

Reference 23

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:15.973885Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 661d7a26-7acf-4570-b022-dc295f2eff51 · outbound

This paper cites Great, Now Write an Article About That: The Crescendo Multi-Turn LLM Jailbreak Attack.

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs Great, Now Write an Article About That: The Crescendo Multi-Turn LLM Jailbreak Attack

Reference 24

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:15.978011Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e620d36b-1e62-46d8-af9a-40cd00a1cf4b · outbound

This paper cites Thakkar, M.; Fournier, Q.; Riemer, M.; Chen, P.-Y.; Zouaq, A.; Das, P.; and Chandar, S.

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs Thakkar, M.; Fournier, Q.; Riemer, M.; Chen, P.-Y.; Zouaq, A.; Das, P.; and Chandar, S

Reference 25

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:15.981197Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 0aed9cd0-699d-49f4-a2b1-3cbe23fd347a · outbound

This paper cites Combining Domain and Alignment Vectors to Achieve Better Knowledge-Safety Trade-offs in LLMs.

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs Combining Domain and Alignment Vectors to Achieve Better Knowledge-Safety Trade-offs in LLMs

Reference 26

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:15.984176Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation df4dee33-b099-436b-821c-ee82e644ba1c · outbound

This paper cites Assessing the Brittleness of Safety Alignment via Pruning and Low-Rank Modifications.

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs Assessing the Brittleness of Safety Alignment via Pruning and Low-Rank Modifications

Reference 28

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:15.991367Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a0068464-d3d9-4cf4-9eef-4d141eb2f968 · outbound

This paper cites InIEEE Symposium on Security and Privacy (S&P).

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs InIEEE Symposium on Security and Privacy (S&P)

Reference 30

Resolution: verified exact; raw_fallback, observed 2026-08-14T04:46:16.282205Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation e8a0e393-1d01-47bf-9d95-30e914561159 · outbound

This paper cites Bag of Tricks: Benchmarking of Jailbreak Attacks on LLMs.

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs Bag of Tricks: Benchmarking of Jailbreak Attacks on LLMs

Reference 31

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:16.001976Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 48faf8a9-8c10-4345-8027-8a2ed3018bd4 · outbound

This paper cites TIES-Merging: Resolving Interference When Merging Models.

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs TIES-Merging: Resolving Interference When Merging Models

Reference 32

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:16.005471Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation cc05a525-f187-4a51-91da-fc87494fd375 · outbound

This paper cites Mitigating the Backdoor Effect for Multi-Task Model Merging via Safety-Aware Subspace.

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs Mitigating the Backdoor Effect for Multi-Task Model Merging via Safety-Aware Subspace

Reference 33

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:16.009315Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 295ba9dd-4fd7-4a7e-b69e-5d522a441f8d · outbound

This paper cites Language Models are Super Mario: Absorbing Abilities from Homologous Models as a Free Lunch.

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs Language Models are Super Mario: Absorbing Abilities from Homologous Models as a Free Lunch

Reference 34

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:16.012609Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1ac8c5ef-3832-406b-b9f0-0167555c2ab6 · outbound

This paper cites 13 Zhang, J.; He, Y.; Cai, K.; Zhao, H.; Suya, F.; and Tian, Y.

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs 13 Zhang, J.; He, Y.; Cai, K.; Zhao, H.; Suya, F.; and Tian, Y

Reference 35

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:16.015991Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 816b8940-0521-4aa9-8f69-e4b1d337326e · outbound

This paper cites RogueMerge: Robust and Unified Attacks against LLM Model Merging.

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs RogueMerge: Robust and Unified Attacks against LLM Model Merging

Reference 36

Resolution: metadata mismatch; local_arxiv, observed 2026-08-14T04:46:16.084279Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation 98f5e297-da19-442b-ad61-891ff6f9b58f · outbound

This paper cites Instruction-Following Evaluation for Large Language Models.

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs Instruction-Following Evaluation for Large Language Models

Reference 37

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:16.024015Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6c633179-6ea7-4e8f-95ce-bdbd5f951d6a · outbound

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

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs Universal and Transferable Adversarial Attacks on Aligned Language Models

Reference 38

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:16.027663Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b5977ead-c045-42d7-8b35-919c91173192 · outbound

This paper cites On Adaptive Attacks to Adversarial Example Defenses.

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs On Adaptive Attacks to Adversarial Example Defenses

Reference 2020

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:15.987575Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c5912dd9-fde1-4e9a-ae1b-df767c20100c · outbound

This paper cites Evaluating Large Language Models Trained on Code.

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs Evaluating Large Language Models Trained on Code

Reference 2021

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:15.914300Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 18a463a7-77b2-4c50-ba3f-5b93303b948c · outbound

This paper cites Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference time.

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference time

Reference 2022

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:15.995296Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 47639529-db75-489e-bd56-2232d42798dc · outbound

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

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs Jailbreaking Black Box Large Language Models in Twenty Queries

Reference 2023

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:15.910426Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d44097e4-6445-4aeb-bbc8-f3be439fbf4f · outbound

This paper cites Refusal in Language Models Is Mediated by a Single Direction.

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs Refusal in Language Models Is Mediated by a Single Direction

Reference 2024

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:15.897691Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 51554e35-6a93-4cf9-9a88-4642ef2b828a · outbound

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

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs Jailbreaking Leading Safety-Aligned LLMs with Simple Adaptive Attacks

Reference 2025

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:15.893135Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 22524470-4f0d-4b99-b3eb-a6973346ebee · outbound

This paper cites Djuhera, A.; Kadhe, S.

When Skills Meet Safety: Benchmarking and Characterizing the Adaptive Jailbreak Robustness of Skill-Merged LLMs Djuhera, A.; Kadhe, S

Reference 2026

Resolution: unresolved; no resolver link, observed 2026-08-14T04:46:15.922065Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Pith citing papers

No inbound Pith citation observations are available.