Paper Citation Record · LEDGER

Safety Alignment Should Be Made More Than Just a Few Tokens Deep

As of 5 August 2026, this Paper Citation Record lists 0 of 0 outbound references and 42 inbound Pith citation observations for arXiv:2406.05946.

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

pith.paper-citation-record.v1
2406.05946 v1

Coverage vector

measured 0 of 0 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links

measured 42 of 42 standing notices

One-hop event checks from named stored sources.

Source: scholarly_work_events, retraction_status_cache, observed 2026-08-05T06:32:48.257954+00:00

measured 42 of 42 inbound itemization

Pith citing papers itemized under the disclosed page cap.

Source: paper_references, paper_reference_links, observed 2026-08-04T09:47:24.461017Z

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

0 of 0 outbound references displayed

verified exact0
verified fuzzy0
unresolved0
parse uncertain0
malformed identifier0
metadata mismatch0

External citation measurements

4
arxiv_reference, observed 2026-08-05T02:28:24.338817Z

Outbound references

No outbound reference observations are available for this paper version.

Pith citing papers

Observation 18e07020-4633-4cca-ac77-ddc13a509810 · inbound

Harmful Fine-tuning Attacks and Defenses for Large Language Models: A Survey cites this paper.

Harmful Fine-tuning Attacks and Defenses for Large Language Models: A Survey Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 122

Resolution: verified exact; arxiv_id, observed 2026-05-23T20:58:26.408949Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-05T06:32:48.257954+00:00.

Observation 0a3315e6-5034-416a-a3a2-574c9ae1d9e1 · inbound

LLM-Safety Evaluations Lack Robustness cites this paper.

LLM-Safety Evaluations Lack Robustness Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 44

Resolution: verified exact; arxiv_id, observed 2026-05-23T01:27:21.321187Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-05T06:32:48.257954+00:00.

Observation 69257a21-2ff3-458d-9a67-ac2adfc991f8 · inbound

Towards provable probabilistic safety for scalable embodied AI systems cites this paper.

Towards provable probabilistic safety for scalable embodied AI systems Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 138

Resolution: verified exact; arxiv_id, observed 2026-05-19T11:02:15.176084Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-05T06:32:48.257954+00:00.

Observation 31a3c0fc-9898-45f0-8b3b-81da89cdccf5 · inbound

ReasoningGuard: Safeguarding Large Reasoning Models with Inference-time Safety Aha Moments cites this paper.

ReasoningGuard: Safeguarding Large Reasoning Models with Inference-time Safety Aha Moments Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 42

Resolution: verified exact; arxiv_id, observed 2026-05-19T01:02:54.865061Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-05T06:32:48.257954+00:00.

Observation b18694bb-f4ed-470d-95df-23f3ff646588 · inbound

Attention Misses Visual Risk: Risk-Adaptive Steering for Multimodal Safety Alignment cites this paper.

Attention Misses Visual Risk: Risk-Adaptive Steering for Multimodal Safety Alignment Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 15

Resolution: unresolved; no resolver link, observed 2026-08-04T09:47:24.461017Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9a85dacc-f99a-4773-9cb3-851478a24a55 · inbound

Reasoning Up the Instruction Ladder for Controllable Language Models cites this paper.

Reasoning Up the Instruction Ladder for Controllable Language Models Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 27

Resolution: unresolved; no resolver link, observed 2026-08-04T07:07:54.996168Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation cfbb6043-d7a0-4edb-9bf6-cf0e2bbfa64c · inbound

SCOUT: A Defense Against Data Poisoning Attacks in Fine-Tuned Language Models cites this paper.

SCOUT: A Defense Against Data Poisoning Attacks in Fine-Tuned Language Models Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 26

Resolution: verified exact; arxiv_id, observed 2026-05-16T23:28:40.713873Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-05T06:32:48.257954+00:00.

Observation 796e4efe-4123-4436-9ed9-85fff8302d09 · inbound

Revisiting Robustness for LLM Safety Alignment via Selective Geometry Control cites this paper.

Revisiting Robustness for LLM Safety Alignment via Selective Geometry Control Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 20

Resolution: verified exact; arxiv_id, observed 2026-05-22T11:21:29.072310Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-05T06:32:48.257954+00:00.

Observation c534c6af-9e58-40e0-bbba-33d68b94a85d · inbound

Robust Policy Optimization to Prevent Catastrophic Forgetting cites this paper.

Robust Policy Optimization to Prevent Catastrophic Forgetting Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 39

Resolution: verified exact; arxiv_id, observed 2026-05-16T05:37:24.169526Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-05T06:32:48.257954+00:00.

Observation aefd7fba-2534-46b6-8ec3-068284866237 · inbound

IatroBench: Pre-Registered Evidence of Iatrogenic Harm from AI Safety Measures cites this paper.

IatroBench: Pre-Registered Evidence of Iatrogenic Harm from AI Safety Measures Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 28

Resolution: verified exact; arxiv_id, observed 2026-05-11T00:35:52.662413Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-05T06:32:48.257954+00:00.

Observation 3731c79d-7bc6-4d28-9d1f-c74def9e77ad · inbound

IatroBench: Pre-Registered Evidence of Iatrogenic Harm from AI Safety Measures cites this paper.

IatroBench: Pre-Registered Evidence of Iatrogenic Harm from AI Safety Measures Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 28

Resolution: unresolved; no resolver link, observed 2026-07-13T00:19:33.861692Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a2dfbc05-f563-47c7-aa30-380075ba452c · inbound

Are GUI Agents Focused Enough? Automated Distraction via Semantic-level UI Element Injection cites this paper.

Are GUI Agents Focused Enough? Automated Distraction via Semantic-level UI Element Injection Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 37

Resolution: metadata mismatch; arxiv_id, observed 2026-05-11T00:30:52.978026Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-05T06:32:48.257954+00:00.

Observation c548783a-0369-45e7-9738-d533958b6ae5 · inbound

Large Language Models Generate Harmful Responses Using a Distinct Mechanism, Shared Across Harm Types cites this paper.

Large Language Models Generate Harmful Responses Using a Distinct Mechanism, Shared Across Harm Types Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 20

Resolution: verified exact; arxiv_id, observed 2026-05-11T07:31:00.406851Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-05T06:32:48.257954+00:00.

Observation 5f398ca0-4b21-4259-9724-775fd701064c · inbound

Latent Instruction Representation Alignment: defending against jailbreaks, backdoors and undesired knowledge in LLMs cites this paper.

Latent Instruction Representation Alignment: defending against jailbreaks, backdoors and undesired knowledge in LLMs Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 30

Resolution: verified exact; arxiv_id, observed 2026-05-11T08:21:00.097938Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-05T06:32:48.257954+00:00.

Observation ccc0fa3b-dfe9-4e8b-b4b9-e7df34205fd6 · inbound

MultiBreak: A Scalable and Diverse Multi-turn Jailbreak Benchmark for Evaluating LLM Safety cites this paper.

MultiBreak: A Scalable and Diverse Multi-turn Jailbreak Benchmark for Evaluating LLM Safety Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 8

Resolution: metadata mismatch; arxiv_id, observed 2026-05-11T09:31:00.975345Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-05T06:32:48.257954+00:00.

Observation a2869571-ad0f-4ee1-8546-bbc2319adc3a · inbound

RouteHijack: Routing-Aware Attack on Mixture-of-Experts LLMs cites this paper.

RouteHijack: Routing-Aware Attack on Mixture-of-Experts LLMs Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 45

Resolution: verified exact; arxiv_id, observed 2026-05-11T15:46:18.735027Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-05T06:32:48.257954+00:00.

Observation af70cf46-cef4-42b9-b766-97bcd16061eb · inbound

Internalizing Safety Understanding in Large Reasoning Models via Verification cites this paper.

Internalizing Safety Understanding in Large Reasoning Models via Verification Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 15

Resolution: verified exact; arxiv_id, observed 2026-05-12T01:51:14.234952Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-05T06:32:48.257954+00:00.

Observation f8c0cbb3-a741-458e-9e23-509d52f38403 · inbound

REALISTA: Realistic Latent Adversarial Attacks that Elicit LLM Hallucinations cites this paper.

REALISTA: Realistic Latent Adversarial Attacks that Elicit LLM Hallucinations Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 191

Resolution: verified exact; arxiv_id, observed 2026-05-14T20:19:26.730208Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-05T06:32:48.257954+00:00.

Observation b8910e1c-87c9-4d1a-ab7a-1ba54351b815 · inbound

Reducing the Safety Tax in LLM Safety Alignment with On-Policy Self-Distillation cites this paper.

Reducing the Safety Tax in LLM Safety Alignment with On-Policy Self-Distillation Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 13

Resolution: metadata mismatch; arxiv_id, observed 2026-05-19T16:37:39.886052Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-05T06:32:48.257954+00:00.

Observation 50336ac5-2974-406f-ac2d-16be30884918 · inbound

Ablating Safety: Mechanisms for Removing Alignment in Language Models for Security Applications cites this paper.

Ablating Safety: Mechanisms for Removing Alignment in Language Models for Security Applications Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 32

Resolution: verified exact; arxiv_id, observed 2026-05-19T23:32:52.534733Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-05T06:32:48.257954+00:00.

Observation 32b70579-c676-4e14-9cb8-9813176209ca · inbound

ParaVT: Taming the Tool Prior Paradox for Parallel Tool Use in Agentic Video Reinforcement Learning cites this paper.

ParaVT: Taming the Tool Prior Paradox for Parallel Tool Use in Agentic Video Reinforcement Learning Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 23

Resolution: verified exact; arxiv_id, observed 2026-05-21T07:34:02.553168Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-05T06:32:48.257954+00:00.

Observation 0c86dbcc-5cd4-4e1c-85ed-99cd3007d537 · inbound

ParaVT: Taming the Tool Prior Paradox for Parallel Tool Use in Agentic Video Reinforcement Learning cites this paper.

ParaVT: Taming the Tool Prior Paradox for Parallel Tool Use in Agentic Video Reinforcement Learning Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 23

Resolution: verified exact; arxiv_id, observed 2026-05-22T09:01:19.405588Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-05T06:32:48.257954+00:00.

Observation 45b8de59-d327-4b5c-a66e-34520b9d6e45 · inbound

REFLECTOR: Internalizing Step-wise Reflection against Indirect Jailbreak cites this paper.

REFLECTOR: Internalizing Step-wise Reflection against Indirect Jailbreak Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 13

Resolution: metadata mismatch; arxiv_id, observed 2026-05-21T06:19:41.984616Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-05T06:32:48.257954+00:00.

Observation e74004bf-4bd9-4474-990e-04ed45a96364 · inbound

Detecting Unfaithful Chain-of-Thought via Circuit-Guided Internal-External Discrepancy cites this paper.

Detecting Unfaithful Chain-of-Thought via Circuit-Guided Internal-External Discrepancy Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 38

Resolution: verified exact; arxiv_id, observed 2026-06-29T21:53:59.549110Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-05T06:32:48.257954+00:00.

Observation eeb1c9dd-8132-4ef8-9bf8-8e403c29f35d · inbound

MESA: Improving MoE Safety Alignment via Decentralized Expertise cites this paper.

MESA: Improving MoE Safety Alignment via Decentralized Expertise Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 17

Resolution: metadata mismatch; arxiv_id, observed 2026-06-28T19:52:35.291030Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-05T06:32:48.257954+00:00.

Observation 54d52469-bcff-45c6-be6b-ec2b969b7c84 · inbound

MENTIS: What Belief Changes Under Alignment? Measuring Multi-Scale Latent Torsion in Language Models cites this paper.

MENTIS: What Belief Changes Under Alignment? Measuring Multi-Scale Latent Torsion in Language Models Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 6

Resolution: metadata mismatch; arxiv_id, observed 2026-07-01T21:16:13.704818Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-05T06:32:48.257954+00:00.

Observation 40e2d7f2-51b7-4498-9ae6-418d195d0f72 · inbound

MaskForge: Structure-Aware Adaptive Attacks for Jailbreaking Diffusion Large Language Models cites this paper.

MaskForge: Structure-Aware Adaptive Attacks for Jailbreaking Diffusion Large Language Models Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 28

Resolution: verified exact; arxiv_id, observed 2026-07-01T23:56:24.322243Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-05T06:32:48.257954+00:00.

Observation d974c386-bb2b-494d-bd0a-1181b7910d2b · inbound

When Behavioral Safety Evaluation Fails: A Representation-Level Perspective cites this paper.

When Behavioral Safety Evaluation Fails: A Representation-Level Perspective Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 9

Resolution: metadata mismatch; arxiv_id, observed 2026-07-02T20:57:23.042879Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-05T06:32:48.257954+00:00.

Observation db63a811-404a-41be-84c2-f94e3316018c · inbound

Breaking the Solver Bottleneck: Training Task Generators at the Learnable Frontier cites this paper.

Breaking the Solver Bottleneck: Training Task Generators at the Learnable Frontier Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 105

Resolution: metadata mismatch; arxiv_id, observed 2026-07-03T08:57:48.136806Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-05T06:32:48.257954+00:00.

Observation 3509ee17-abb0-4704-ad67-20241ed6968e · inbound

Understanding and Mitigating Prompt Leaking Attacks in Real-World LLM-Based Applications cites this paper.

Understanding and Mitigating Prompt Leaking Attacks in Real-World LLM-Based Applications Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 54

Resolution: verified exact; arxiv_id, observed 2026-07-04T00:59:20.668111Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-05T06:32:48.257954+00:00.

Observation d4d13048-04c3-4378-bfaa-27f6348b12c1 · inbound

Forget, Anticipate and Adapt: Test Time Training for Long Videos cites this paper.

Forget, Anticipate and Adapt: Test Time Training for Long Videos Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 43

Resolution: verified exact; arxiv_id, observed 2026-07-04T12:59:52.982067Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-05T06:32:48.257954+00:00.

Observation 822981c2-9a40-4540-a0b8-1fb41754b31a · inbound

Forget, Anticipate and Adapt: Test Time Training for Long Videos cites this paper.

Forget, Anticipate and Adapt: Test Time Training for Long Videos Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 43

Resolution: verified exact; arxiv_id, observed 2026-06-30T11:54:38.971589Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-05T06:32:48.257954+00:00.

Observation 72b76032-4b80-44a6-9452-18ff921c8d81 · inbound

Forget, Anticipate and Adapt: Test Time Training for Long Videos cites this paper.

Forget, Anticipate and Adapt: Test Time Training for Long Videos Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 43

Resolution: unresolved; no resolver link, observed 2026-07-14T17:16:44.626166Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 08176c57-f6af-417e-bb42-155045e9ae68 · inbound

Defending Against Harmful Supervision Hidden in Benign Samples cites this paper.

Defending Against Harmful Supervision Hidden in Benign Samples Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 3

Resolution: metadata mismatch; arxiv_id, observed 2026-06-30T14:24:45.328038Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-05T06:32:48.257954+00:00.

Observation 52b6147b-d107-47f2-99c0-493177db0293 · inbound

Breaking Safety at the Token Boundary: How BPE Tokenization Creates Exploitable Gaps in LLM Alignment cites this paper.

Breaking Safety at the Token Boundary: How BPE Tokenization Creates Exploitable Gaps in LLM Alignment Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 9

Resolution: verified exact; arxiv_id, observed 2026-07-04T01:19:20.146959Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-05T06:32:48.257954+00:00.

Observation 13b80043-0285-4561-8a97-cda8b8ec98cb · inbound

Transplanting, inverting, and preventing a misalignment persona: method-conditional emergent misalignment in Qwen2.5 cites this paper.

Transplanting, inverting, and preventing a misalignment persona: method-conditional emergent misalignment in Qwen2.5 Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 12

Resolution: unresolved; no resolver link, observed 2026-07-11T18:20:40.287006Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8b0e5569-2443-405d-8e81-3555d29cf70a · inbound

Pretraining Curricula Enable Selective Fine-tuning cites this paper.

Pretraining Curricula Enable Selective Fine-tuning Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 48

Resolution: unresolved; no resolver link, observed 2026-07-11T12:36:24.747752Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 3df96879-e39b-4505-8a87-b1cbe822d585 · inbound

An Emergent Mirage: Is Emergent Misalignment and Realignment Indeed a Robust Phenomenon? cites this paper.

An Emergent Mirage: Is Emergent Misalignment and Realignment Indeed a Robust Phenomenon? Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 39

Resolution: unresolved; no resolver link, observed 2026-07-13T00:42:24.432562Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 54e89284-a98d-4c11-8add-40007f5d5d99 · inbound

Breaking Refusal in the First Half: A Mechanistic Study of the Prefill Jailbreak cites this paper.

Breaking Refusal in the First Half: A Mechanistic Study of the Prefill Jailbreak Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 47

Resolution: unresolved; no resolver link, observed 2026-08-02T06:30:56.251349Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4c5b5df1-bb7b-4950-8fc5-89572e115e3b · inbound

Normalized Rewards for Preference Optimization cites this paper.

Normalized Rewards for Preference Optimization Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 14

Resolution: unresolved; no resolver link, observed 2026-08-02T10:01:58.569910Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e94f6a3c-d03e-4e58-8f17-d7a4b741ab86 · inbound

How Jailbreak Attacks Inform Safety Alignment: A Defender-Centric, Shapley-Based Evaluation of Jailbreak Contributions cites this paper.

How Jailbreak Attacks Inform Safety Alignment: A Defender-Centric, Shapley-Based Evaluation of Jailbreak Contributions Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 57

Resolution: unresolved; no resolver link, observed 2026-08-01T18:54:57.776258Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 18bb393f-b765-4321-a81a-61f84477da3b · inbound

Reason Before You Retrieve: Agentic Planning for Multi-modal RAG cites this paper.

Reason Before You Retrieve: Agentic Planning for Multi-modal RAG Safety Alignment Should Be Made More Than Just a Few Tokens Deep

Reference 58

Resolution: unresolved; no resolver link, observed 2026-08-02T10:20:54.536465Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.