Typed states for the displayed outbound observations.
Source: paper_references, paper_reference_links, observed 2026-08-06T15:01:47.854894Z
Paper Citation Record · LEDGER
The Geometry of Harmfulness in LLMs through Subconcept Probing
As of 20 August 2026, this Paper Citation Record lists 28 of 28 outbound references and 2 inbound Pith citation observations for arXiv:2507.21141.
A citation records a reference. It does not transfer a finding from one paper to another.
Coverage vector
One-hop event checks from named stored sources.
Source: scholarly_work_events, retraction_status_cache, observed 2026-08-19T06:32:44.657259+00:00
Pith citing papers itemized under the disclosed page cap.
Source: paper_references, paper_reference_links, observed 2026-08-02T07:09:32.736304Z
A source-named dated measurement, never combined with another source.
Source: arxiv_reference, observed 2026-05-14T21:38:00.071469Z
Reference resolution
28 of 28 outbound references displayed
- verified exact0
- verified fuzzy4
- unresolved24
- parse uncertain0
- malformed identifier0
- metadata mismatch0
External citation measurements
No source-named external measurement is stored.
Outbound references
Observation ecb6aa75-1eaf-4e6c-89a3-701ad29a7dff · outbound
This paper cites On the Opportunities and Risks of Foundation Models.
The Geometry of Harmfulness in LLMs through Subconcept Probing On the Opportunities and Risks of Foundation Models
Reference 3
- Resolution
- unresolved
- no resolver link, observed 2026-08-06T15:01:47.765176Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 82c5c459-c5e7-4430-96b4-4da7071c1999 · outbound
This paper cites Safety-Aware Fine-Tuning of Large Language Models.
The Geometry of Harmfulness in LLMs through Subconcept Probing Safety-Aware Fine-Tuning of Large Language Models
Reference 5
- Resolution
- unresolved
- no resolver link, observed 2026-08-06T15:01:47.772974Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation f360be14-326c-43cc-8947-f019ae1cba39 · outbound
This paper cites Training Verifiers to Solve Math Word Problems.
The Geometry of Harmfulness in LLMs through Subconcept Probing Training Verifiers to Solve Math Word Problems
Reference 6
- Resolution
- unresolved
- no resolver link, observed 2026-08-06T15:01:47.776571Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 7fb6beaf-c30a-4a6e-a4a4-0d5bf79bceec · outbound
This paper cites The Llama 3 Herd of Models.
The Geometry of Harmfulness in LLMs through Subconcept Probing The Llama 3 Herd of Models
Reference 10
- Resolution
- unresolved
- no resolver link, observed 2026-08-06T15:01:47.791537Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation a528a80c-c0b7-4f4d-9440-a7b03bb7d901 · outbound
This paper cites Safedpo: A simple approach to direct preference optimization with enhanced safety.
The Geometry of Harmfulness in LLMs through Subconcept Probing Safedpo: A simple approach to direct preference optimization with enhanced safety
Reference 13
- Resolution
- unresolved
- no resolver link, observed 2026-08-06T15:01:47.801834Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 6b0343f1-d2dc-41e9-9564-a4d601b82724 · outbound
This paper cites AutoDAN: Generating Stealthy Jailbreak Prompts on Aligned Large Language Models.
The Geometry of Harmfulness in LLMs through Subconcept Probing AutoDAN: Generating Stealthy Jailbreak Prompts on Aligned Large Language Models
Reference 14
- Resolution
- unresolved
- no resolver link, observed 2026-08-06T15:01:47.805088Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation a1939382-08e1-4005-bcf6-ae0c11ce47c8 · outbound
This paper cites Enhancing LLM Safety via Constrained Direct Preference Optimization.
The Geometry of Harmfulness in LLMs through Subconcept Probing Enhancing LLM Safety via Constrained Direct Preference Optimization
Reference 15
- Resolution
- unresolved
- no resolver link, observed 2026-08-06T15:01:47.808584Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation ff0ef252-7b16-4373-8618-3decc99bb851 · outbound
This paper cites The Geometry of Truth: Emergent Linear Structure in Large Language Model Representations of True/False Datasets.
The Geometry of Harmfulness in LLMs through Subconcept Probing The Geometry of Truth: Emergent Linear Structure in Large Language Model Representations of True/False Datasets
Reference 16
- Resolution
- unresolved
- no resolver link, observed 2026-08-06T15:01:47.812006Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation a681fd9b-fb00-472c-b52d-078489649203 · outbound
This paper cites Sparse Feature Circuits: Discovering and Editing Interpretable Causal Graphs in Language Models.
The Geometry of Harmfulness in LLMs through Subconcept Probing Sparse Feature Circuits: Discovering and Editing Interpretable Causal Graphs in Language Models
Reference 17
- Resolution
- unresolved
- no resolver link, observed 2026-08-06T15:01:47.815767Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 509a0729-c1e0-46ab-a923-9a4e81bc055d · outbound
This paper cites Emergent Linear Representations in World Models of Self-Supervised Sequence Models.
The Geometry of Harmfulness in LLMs through Subconcept Probing Emergent Linear Representations in World Models of Self-Supervised Sequence Models
Reference 18
- Resolution
- unresolved
- no resolver link, observed 2026-08-06T15:01:47.819681Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 70ffcc5d-2c1c-4592-ac8a-bbbb34600921 · outbound
This paper cites The Hidden Dimensions of LLM Alignment: A Multi-Dimensional Analysis of Orthogonal Safety Directions.
The Geometry of Harmfulness in LLMs through Subconcept Probing The Hidden Dimensions of LLM Alignment: A Multi-Dimensional Analysis of Orthogonal Safety Directions
Reference 19
- Resolution
- unresolved
- no resolver link, observed 2026-08-06T15:01:47.822986Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 0a8fb2e9-746c-4bd9-808e-5dc127b0f233 · outbound
This paper cites Interpretable steering of large language models with feature guided activation additions.
The Geometry of Harmfulness in LLMs through Subconcept Probing Interpretable steering of large language models with feature guided activation additions
Reference 20
- Resolution
- verified fuzzy
- raw_fallback, observed 2026-08-06T15:01:48.208051Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.
Observation a850eefb-8758-4c6a-934d-26aa0304363f · outbound
This paper cites Linear Representations of Sentiment in Large Language Models.
The Geometry of Harmfulness in LLMs through Subconcept Probing Linear Representations of Sentiment in Large Language Models
Reference 21
- Resolution
- unresolved
- no resolver link, observed 2026-08-06T15:01:47.830115Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 4ad54f81-b217-4a3e-b17c-67d477fd1619 · outbound
This paper cites Steering Language Models With Activation Engineering.
The Geometry of Harmfulness in LLMs through Subconcept Probing Steering Language Models With Activation Engineering
Reference 22
- Resolution
- unresolved
- no resolver link, observed 2026-08-06T15:01:47.833410Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 47b0c430-ee7b-49a7-acd6-18f093ce0bce · outbound
This paper cites pyvene: A library for understanding and improving PyTorch models via interventions.
The Geometry of Harmfulness in LLMs through Subconcept Probing pyvene: A library for understanding and improving PyTorch models via interventions
Reference 23
- Resolution
- verified fuzzy
- raw_fallback, observed 2026-08-06T15:01:48.198757Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.
Observation 3407a27e-cdb7-4591-9266-9b563e4e3180 · outbound
This paper cites URL https://aclanthology.org/2024.naacl-demo.
The Geometry of Harmfulness in LLMs through Subconcept Probing URL https://aclanthology.org/2024.naacl-demo
Reference 24
- Resolution
- verified fuzzy
- raw_fallback, observed 2026-08-06T15:01:48.189182Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.
Observation ea24d6d9-d573-408f-b00e-5409aa37c140 · outbound
This paper cites Qwen2 Technical Report.
The Geometry of Harmfulness in LLMs through Subconcept Probing Qwen2 Technical Report
Reference 25
- Resolution
- unresolved
- no resolver link, observed 2026-08-06T15:01:47.843887Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation f5c4d94f-9caa-4c5e-8b5b-e722938ad976 · outbound
This paper cites From Directions to Cones: Exploring Multidimensional Representations of Propositional Facts in LLMs.
The Geometry of Harmfulness in LLMs through Subconcept Probing From Directions to Cones: Exploring Multidimensional Representations of Propositional Facts in LLMs
Reference 26
- Resolution
- unresolved
- no resolver link, observed 2026-08-06T15:01:47.847228Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation e59f07d6-563b-4f03-b42d-8bf5f66c7b67 · outbound
This paper cites Under review.
The Geometry of Harmfulness in LLMs through Subconcept Probing Under review
Reference 27
- Resolution
- verified fuzzy
- raw_fallback, observed 2026-08-06T15:01:48.177822Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.
Observation 323d869c-b384-4e58-b0b8-df3e41f18161 · outbound
This paper cites an unresolved cited work.
The Geometry of Harmfulness in LLMs through Subconcept Probing Unresolved cited work
Reference 28
- Resolution
- unresolved
- raw_fallback, observed 2026-08-06T15:01:48.167501Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.
Observation 4a5a077b-c308-4fea-820a-662d0f93c565 · outbound
This paper cites doi: https:// doi.org/10.1016/S0031-3203(96)00142-2.
The Geometry of Harmfulness in LLMs through Subconcept Probing doi: https:// doi.org/10.1016/S0031-3203(96)00142-2
Reference 1997
- Resolution
- unresolved
- no resolver link, observed 2026-08-06T15:01:47.769206Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 580723f0-2f20-47b0-897a-9660e8d9a8e0 · outbound
This paper cites Refusal Behavior in Large Language Models: A Nonlinear Perspective.
The Geometry of Harmfulness in LLMs through Subconcept Probing Refusal Behavior in Large Language Models: A Nonlinear Perspective
Reference 2002
- Resolution
- unresolved
- no resolver link, observed 2026-08-06T15:01:47.794774Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 15050118-aeec-4901-9351-987148f61d04 · outbound
This paper cites emnlp-main.273.
The Geometry of Harmfulness in LLMs through Subconcept Probing emnlp-main.273
Reference 2020
- Resolution
- unresolved
- no resolver link, observed 2026-08-06T15:01:47.787746Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 2ad56729-0f6b-43cc-88de-f5f7a594dbb2 · outbound
This paper cites Toy Models of Superposition.
The Geometry of Harmfulness in LLMs through Subconcept Probing Toy Models of Superposition
Reference 2021
- Resolution
- unresolved
- no resolver link, observed 2026-08-06T15:01:47.780339Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 266f749f-500e-42c1-9e67-1ad2a6323e5f · outbound
This paper cites an unresolved cited work.
The Geometry of Harmfulness in LLMs through Subconcept Probing Unresolved cited work
Reference 2022
- Resolution
- unresolved
- raw_fallback, observed 2026-08-06T15:01:48.218134Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.
Observation 2b54dfec-7589-47ba-96ab-80ad776a304e · outbound
This paper cites Refusal in Language Models Is Mediated by a Single Direction.
The Geometry of Harmfulness in LLMs through Subconcept Probing Refusal in Language Models Is Mediated by a Single Direction
Reference 2023
- Resolution
- unresolved
- no resolver link, observed 2026-08-06T15:01:47.757141Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 9ab45db4-4393-4a33-b972-58749b0b829a · outbound
This paper cites On the dangers of stochastic parrots: Can language models be too big? In Proceedings of the 2021 ACM conference on fairness, accountability, and transparency, pp.
The Geometry of Harmfulness in LLMs through Subconcept Probing On the dangers of stochastic parrots: Can language models be too big? In Proceedings of the 2021 ACM conference on fairness, accountability, and transparency, pp
Reference 2024
- Resolution
- unresolved
- no resolver link, observed 2026-08-06T15:01:47.761453Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Observation 89355c50-affb-4354-b6b5-88a10c87a02d · outbound
This paper cites On the Origins of Linear Representations in Large Language Models.
The Geometry of Harmfulness in LLMs through Subconcept Probing On the Origins of Linear Representations in Large Language Models
Reference 2025
- Resolution
- unresolved
- no resolver link, observed 2026-08-06T15:01:47.798462Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.
Pith citing papers
Observation 05b043eb-10a1-4ebf-9169-085a654c8545 · inbound
Before the Last Token: Diagnosing Final-Token Safety Probe Failures cites this paper.
Before the Last Token: Diagnosing Final-Token Safety Probe Failures The Geometry of Harmfulness in LLMs through Subconcept Probing
Reference 11
- Resolution
- verified exact
- arxiv_id, observed 2026-05-14T21:38:00.073909Z
Source-reported events for the cited work
No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.
Observation 28a2ee74-b0c1-4dc9-a21e-8d16199d1ba7 · inbound
The Entanglement Wall: Activation-Space Probes as Risk Detectors, Not Context Adjudicators cites this paper.
The Entanglement Wall: Activation-Space Probes as Risk Detectors, Not Context Adjudicators The Geometry of Harmfulness in LLMs through Subconcept Probing
Reference 28
- Resolution
- unresolved
- no resolver link, observed 2026-08-02T07:09:32.736304Z
Source-reported events for the cited work
Unavailable: canonical work link unavailable.