Paper Citation Record · LEDGER

What AI evaluations for preventing catastrophic risks can and cannot do

As of 14 August 2026, this Paper Citation Record lists 29 of 29 outbound references and 3 inbound Pith citation observations for arXiv:2412.08653.

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

pith.paper-citation-record.v1
2412.08653 v1

Coverage vector

measured 29 of 29 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-08-12T11:56:21.102873Z

measured 32 of 32 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 3 of 3 inbound itemization

Pith citing papers itemized under the disclosed page cap.

Source: paper_references, paper_reference_links, observed 2026-06-30T09:01:07.210356Z

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

29 of 29 outbound references displayed

verified exact0
verified fuzzy11
unresolved17
parse uncertain1
malformed identifier0
metadata mismatch0

External citation measurements

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

Outbound references

Observation 7468784d-5e67-48ad-aee4-9c0c36b67029 · outbound

This paper cites Declare and Justify: Explicit assumptions in AI evaluations are necessary for effective regulation.

What AI evaluations for preventing catastrophic risks can and cannot do Declare and Justify: Explicit assumptions in AI evaluations are necessary for effective regulation

Reference 1

Resolution: unresolved; no resolver link, observed 2026-08-12T11:56:20.959799Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4c07cfaa-96fd-4d42-82f1-afecfaf2b215 · outbound

This paper cites Safety Cases: How to Justify the Safety of Advanced AI Systems.

What AI evaluations for preventing catastrophic risks can and cannot do Safety Cases: How to Justify the Safety of Advanced AI Systems

Reference 2

Resolution: unresolved; no resolver link, observed 2026-08-12T11:56:20.965769Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation bc3b5182-caf2-481a-8daa-223e31c20c3e · outbound

This paper cites Safety case template for frontier AI: A cyber inability argument.

What AI evaluations for preventing catastrophic risks can and cannot do Safety case template for frontier AI: A cyber inability argument

Reference 3

Resolution: unresolved; no resolver link, observed 2026-08-12T11:56:20.970943Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c19967bf-392a-4674-a7af-9dbf3aee01c5 · outbound

This paper cites Anthropic’s Responsible Scaling Policy Version 1.0, 2023.

What AI evaluations for preventing catastrophic risks can and cannot do Anthropic’s Responsible Scaling Policy Version 1.0, 2023

Reference 4

Resolution: verified fuzzy; raw_fallback, observed 2026-08-12T11:56:21.577496Z

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 817b2439-fa05-4b40-b1e5-e808212ae3f6 · outbound

This paper cites Preparedness Framework (Beta), 2023.

What AI evaluations for preventing catastrophic risks can and cannot do Preparedness Framework (Beta), 2023

Reference 5

Resolution: verified fuzzy; raw_fallback, observed 2026-08-12T11:56:21.562724Z

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 e81ae6be-27ad-4d7b-9c40-76a12851e49a · outbound

This paper cites Frontier Safety Framework, 2024.

What AI evaluations for preventing catastrophic risks can and cannot do Frontier Safety Framework, 2024

Reference 6

Resolution: verified fuzzy; raw_fallback, observed 2026-08-12T11:56:21.546647Z

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 d7d3fc15-ed60-4b0e-984f-dca0c53661e7 · outbound

This paper cites Evaluating Frontier Models for Dangerous Capabilities.

What AI evaluations for preventing catastrophic risks can and cannot do Evaluating Frontier Models for Dangerous Capabilities

Reference 7

Resolution: unresolved; no resolver link, observed 2026-08-12T11:56:20.991463Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ba8390aa-6224-4285-90d1-2c691ef4fdf2 · outbound

This paper cites LLM Agents can Autonomously Hack Websites.

What AI evaluations for preventing catastrophic risks can and cannot do LLM Agents can Autonomously Hack Websites

Reference 8

Resolution: unresolved; no resolver link, observed 2026-08-12T11:56:20.996657Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e9a0ca4c-7e19-4ec5-af0f-07ede74845fd · outbound

This paper cites LLM Agents can Autonomously Exploit One-day Vulnerabilities.

What AI evaluations for preventing catastrophic risks can and cannot do LLM Agents can Autonomously Exploit One-day Vulnerabilities

Reference 9

Resolution: unresolved; no resolver link, observed 2026-08-12T11:56:21.001715Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8630ca99-e29e-44ea-960c-c5679d8e56af · outbound

This paper cites Teams of LLM Agents can Exploit Zero-Day Vulnerabilities.

What AI evaluations for preventing catastrophic risks can and cannot do Teams of LLM Agents can Exploit Zero-Day Vulnerabilities

Reference 10

Resolution: unresolved; no resolver link, observed 2026-08-12T11:56:21.006726Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2d3b758c-01ec-4f00-8584-2d1dbeda2f49 · outbound

This paper cites On the Conversational Persuasiveness of Large Language Models: A Randomized Controlled Trial.

What AI evaluations for preventing catastrophic risks can and cannot do On the Conversational Persuasiveness of Large Language Models: A Randomized Controlled Trial

Reference 11

Resolution: unresolved; no resolver link, observed 2026-08-12T11:56:21.012376Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 3ac8c7fa-0f30-479f-8f82-fc30170e4a39 · outbound

This paper cites an unresolved cited work.

What AI evaluations for preventing catastrophic risks can and cannot do Unresolved cited work

Reference 12

Resolution: unresolved; raw_fallback, observed 2026-08-12T11:56:21.530325Z

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 12a40c67-8ce5-441d-8cc3-e7c4a83cb7fd · outbound

This paper cites Black-Box Access is Insufficient for Rigorous AI Audits.

What AI evaluations for preventing catastrophic risks can and cannot do Black-Box Access is Insufficient for Rigorous AI Audits

Reference 13

Resolution: unresolved; no resolver link, observed 2026-08-12T11:56:21.024293Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 300b0777-7082-439e-ae89-b4811143885e · outbound

This paper cites Number 1.

What AI evaluations for preventing catastrophic risks can and cannot do Number 1

Reference 14

Resolution: verified fuzzy; raw_fallback, observed 2026-08-12T11:56:21.514424Z

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 26f9a1da-deb7-4799-9710-10215a022e79 · outbound

This paper cites Mistral CEO confirms ‘leak’ of new open source AI model nearing GPT-4 performance.

What AI evaluations for preventing catastrophic risks can and cannot do Mistral CEO confirms ‘leak’ of new open source AI model nearing GPT-4 performance

Reference 15

Resolution: verified fuzzy; raw_fallback, observed 2026-08-12T11:56:21.499309Z

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 b269c1b4-c725-4567-b27c-298ddd69fc68 · outbound

This paper cites Coordinated pausing: An evaluation-based coordination scheme for frontier AI developers.

What AI evaluations for preventing catastrophic risks can and cannot do Coordinated pausing: An evaluation-based coordination scheme for frontier AI developers

Reference 16

Resolution: unresolved; no resolver link, observed 2026-08-12T11:56:21.045148Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 32c574b4-ab9c-4204-806c-30cee30ae79d · outbound

This paper cites CyberSecEval 2: A Wide-Ranging Cybersecurity Evaluation Suite for Large Language Models.

What AI evaluations for preventing catastrophic risks can and cannot do CyberSecEval 2: A Wide-Ranging Cybersecurity Evaluation Suite for Large Language Models

Reference 17

Resolution: unresolved; no resolver link, observed 2026-08-12T11:56:21.049719Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9b452253-cdb7-4371-bea3-85ac17dd387d · outbound

This paper cites Project Naptime: Evaluating Offensive Security Capabili- ties of Large Language Models.

What AI evaluations for preventing catastrophic risks can and cannot do Project Naptime: Evaluating Offensive Security Capabili- ties of Large Language Models

Reference 18

Resolution: verified fuzzy; raw_fallback, observed 2026-08-12T11:56:21.467303Z

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 2cdbc31e-6fd0-40f6-9683-586b08b5b7df · outbound

This paper cites AI capabilities can be significantly improved without expensive retraining.

What AI evaluations for preventing catastrophic risks can and cannot do AI capabilities can be significantly improved without expensive retraining

Reference 19

Resolution: unresolved; no resolver link, observed 2026-08-12T11:56:21.059369Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 0b1e790e-94f8-468b-85ec-2729c0a77599 · outbound

This paper cites SWE-bench: Can language models resolve real-world github issues? In The Twelfth International Conference on Learning Representations, 2024.

What AI evaluations for preventing catastrophic risks can and cannot do SWE-bench: Can language models resolve real-world github issues? In The Twelfth International Conference on Learning Representations, 2024

Reference 20

Resolution: unresolved; no resolver link, observed 2026-08-12T11:56:21.064339Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 0be6524a-f8b9-4584-9f82-8045a94b9f74 · outbound

This paper cites SWE-bench leaderboard.

What AI evaluations for preventing catastrophic risks can and cannot do SWE-bench leaderboard

Reference 21

Resolution: verified fuzzy; raw_fallback, observed 2026-08-12T11:56:21.440729Z

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 6b0b18a3-90b8-43a3-bb7f-733cfd2ffecf · outbound

This paper cites GAIA Leaderboard - a Hugging Face Space by gaia-benchmark.

What AI evaluations for preventing catastrophic risks can and cannot do GAIA Leaderboard - a Hugging Face Space by gaia-benchmark

Reference 22

Resolution: verified fuzzy; raw_fallback, observed 2026-08-12T11:56:21.424786Z

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 4975690c-760e-4c98-ae24-983f7d988b7e · outbound

This paper cites HumanEval Benchmark (Code Generation).

What AI evaluations for preventing catastrophic risks can and cannot do HumanEval Benchmark (Code Generation)

Reference 23

Resolution: verified fuzzy; raw_fallback, observed 2026-08-12T11:56:21.408538Z

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 39d36d8a-4b55-4336-922f-1c3e625b3269 · outbound

This paper cites A survey on in-context learning.

What AI evaluations for preventing catastrophic risks can and cannot do A survey on in-context learning

Reference 24

Resolution: unresolved; no resolver link, observed 2026-08-12T11:56:21.083324Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f7345fa4-c59e-4de9-8973-045ab5437f15 · outbound

This paper cites Anthropic’s Responsible Scaling Policy, 2024.

What AI evaluations for preventing catastrophic risks can and cannot do Anthropic’s Responsible Scaling Policy, 2024

Reference 25

Resolution: verified fuzzy; raw_fallback, observed 2026-08-12T11:56:21.382641Z

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 f59e5af6-936f-44b7-9435-3081475a4b5f · outbound

This paper cites We need a Science of Evals.

What AI evaluations for preventing catastrophic risks can and cannot do We need a Science of Evals

Reference 26

Resolution: verified fuzzy; raw_fallback, observed 2026-08-12T11:56:21.365857Z

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 7c19e7bd-3cd7-4c9b-bfa1-4a682fc8b45e · outbound

This paper cites AI Sandbagging: Language Models can Strategically Underperform on Evaluations.

What AI evaluations for preventing catastrophic risks can and cannot do AI Sandbagging: Language Models can Strategically Underperform on Evaluations

Reference 27

Resolution: unresolved; no resolver link, observed 2026-08-12T11:56:21.097843Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7d36241c-6419-43a1-a959-e4f407471c27 · outbound

This paper cites Stress-Testing Capability Elicitation With Password-Locked Models.

What AI evaluations for preventing catastrophic risks can and cannot do Stress-Testing Capability Elicitation With Password-Locked Models

Reference 28

Resolution: unresolved; no resolver link, observed 2026-08-12T11:56:21.102873Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8c2a2164-8318-49f5-bb5a-3d9fa2172135 · outbound

This paper cites an unresolved cited work.

What AI evaluations for preventing catastrophic risks can and cannot do Unresolved cited work

Reference 2024

Resolution: parse uncertain; raw_fallback, observed 2026-08-12T11:56:21.483430Z

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.

Pith citing papers

Observation 8f23a12e-520b-4ff7-b021-a0d8eb82b270 · inbound

From Disclosure to Self-Referential Opacity: Six Dimensions of Strain in Current AI Governance cites this paper.

From Disclosure to Self-Referential Opacity: Six Dimensions of Strain in Current AI Governance What AI evaluations for preventing catastrophic risks can and cannot do

Reference 13

Resolution: verified exact; arxiv_id, observed 2026-05-10T12:00:22.254888Z

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 c9e9d58b-a850-459d-9472-b294307149e7 · inbound

Scaffold Effects on GAIA: A Controlled Comparison cites this paper.

Scaffold Effects on GAIA: A Controlled Comparison What AI evaluations for preventing catastrophic risks can and cannot do

Reference 1

Resolution: verified exact; arxiv_id, observed 2026-07-02T23:07:27.088498Z

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 7629eaf4-63ee-48b1-8620-7afde37db9c6 · inbound

Verifying Restrictions on Frontier AI Research cites this paper.

Verifying Restrictions on Frontier AI Research What AI evaluations for preventing catastrophic risks can and cannot do

Reference 28

Resolution: verified exact; arxiv_id, observed 2026-06-30T09:04:32.211462Z

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.