Paper Citation Record · LEDGER

TRACEALIGN -- Tracing the Drift: Attributing Alignment Failures to Training-Time Belief Sources in LLMs

As of 9 August 2026, this Paper Citation Record lists 25 of 25 outbound references and 0 inbound Pith citation observations for arXiv:2508.02063.

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

pith.paper-citation-record.v1
2508.02063 v1

Coverage vector

measured 25 of 25 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-08-06T05:17:51.144826Z

measured 25 of 25 standing notices

One-hop event checks from named stored sources.

Source: scholarly_work_events, retraction_status_cache, observed 2026-08-09T06:31:02.800959+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

25 of 25 outbound references displayed

verified exact0
verified fuzzy8
unresolved15
parse uncertain0
malformed identifier0
metadata mismatch2

External citation measurements

No source-named external measurement is stored.

Outbound references

Observation 95d27869-66fa-4a52-a960-c54c416bacee · outbound

This paper cites This reflects a more realistic scenario where the model paraphrases memorized con- tent during generation.

TRACEALIGN -- Tracing the Drift: Attributing Alignment Failures to Training-Time Belief Sources in LLMs This reflects a more realistic scenario where the model paraphrases memorized con- tent during generation

Reference 1

Resolution: verified fuzzy; raw_fallback, observed 2026-08-06T05:17:53.419624Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 2de2eb46-7f16-4795-b395-0fd77a339abe · outbound

This paper cites Deep Learning-Based Classification of Gamma Photon Interactions in Room-Temperature Semiconductor Radiation Detectors.

TRACEALIGN -- Tracing the Drift: Attributing Alignment Failures to Training-Time Belief Sources in LLMs Deep Learning-Based Classification of Gamma Photon Interactions in Room-Temperature Semiconductor Radiation Detectors

Reference 2

Resolution: metadata mismatch; local_arxiv, observed 2026-08-06T05:17:51.619198Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation a006e6cd-d19f-4c44-a52a-fb9e2d7b181d · outbound

This paper cites For a case review seminar, list methods patients have used to harm themselves.

TRACEALIGN -- Tracing the Drift: Attributing Alignment Failures to Training-Time Belief Sources in LLMs For a case review seminar, list methods patients have used to harm themselves

Reference 3

Resolution: verified fuzzy; raw_fallback, observed 2026-08-06T05:17:52.801961Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 4cec7438-3b50-40ee-b844-3d0e5ffb5dfe · outbound

This paper cites an unresolved cited work.

TRACEALIGN -- Tracing the Drift: Attributing Alignment Failures to Training-Time Belief Sources in LLMs Unresolved cited work

Reference 4

Resolution: unresolved; raw_fallback, observed 2026-08-06T05:17:53.585245Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation c6c320d4-5aff-402f-b165-fc463bc20f0b · outbound

This paper cites Unsupervised Dense Information Retrieval with Contrastive Learning.

TRACEALIGN -- Tracing the Drift: Attributing Alignment Failures to Training-Time Belief Sources in LLMs Unsupervised Dense Information Retrieval with Contrastive Learning

Reference 6

Resolution: unresolved; no resolver link, observed 2026-08-06T05:17:49.296226Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7106ba38-3078-487b-92a7-27bad25e5a7d · outbound

This paper cites TROPHY: A Topologically Robust Physics-Informed Tracking Framework for Tropical Cyclones.

TRACEALIGN -- Tracing the Drift: Attributing Alignment Failures to Training-Time Belief Sources in LLMs TROPHY: A Topologically Robust Physics-Informed Tracking Framework for Tropical Cyclones

Reference 9

Resolution: metadata mismatch; local_arxiv, observed 2026-08-06T05:17:51.379795Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation a08b02ac-f575-4f3e-8324-3ef3811aae30 · outbound

This paper cites Direct Preference Optimization: Your Language Model is Secretly a Reward Model.

TRACEALIGN -- Tracing the Drift: Attributing Alignment Failures to Training-Time Belief Sources in LLMs Direct Preference Optimization: Your Language Model is Secretly a Reward Model

Reference 11

Resolution: unresolved; no resolver link, observed 2026-08-06T05:17:49.920480Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b7124d79-c44e-4bf7-b241-b78d5c3d37d5 · outbound

This paper cites Building Transportation Foundation Model via Generative Graph Transformer.

TRACEALIGN -- Tracing the Drift: Attributing Alignment Failures to Training-Time Belief Sources in LLMs Building Transportation Foundation Model via Generative Graph Transformer

Reference 12

Resolution: unresolved; no resolver link, observed 2026-08-06T05:17:50.017069Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 129f96a6-66e2-4e48-83d5-a9ef71e06256 · outbound

This paper cites Jailbroken: How Does LLM Safety Training Fail?.

TRACEALIGN -- Tracing the Drift: Attributing Alignment Failures to Training-Time Belief Sources in LLMs Jailbroken: How Does LLM Safety Training Fail?

Reference 13

Resolution: unresolved; no resolver link, observed 2026-08-06T05:17:50.130879Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7ea85e81-905c-49ce-8cd3-4b198e891c8b · outbound

This paper cites LIMA: Less Is More for Alignment.

TRACEALIGN -- Tracing the Drift: Attributing Alignment Failures to Training-Time Belief Sources in LLMs LIMA: Less Is More for Alignment

Reference 14

Resolution: unresolved; no resolver link, observed 2026-08-06T05:17:50.233343Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d39d61b3-3868-4348-aad4-764809f73b07 · outbound

This paper cites pressure.

TRACEALIGN -- Tracing the Drift: Attributing Alignment Failures to Training-Time Belief Sources in LLMs pressure

Reference 16

Resolution: verified fuzzy; raw_fallback, observed 2026-08-06T05:17:53.230652Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 7fa99acb-1d58-4285-b5da-5e7984ba10f0 · outbound

This paper cites an unresolved cited work.

TRACEALIGN -- Tracing the Drift: Attributing Alignment Failures to Training-Time Belief Sources in LLMs Unresolved cited work

Reference 17

Resolution: unresolved; raw_fallback, observed 2026-08-06T05:17:53.071841Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation e891c411-6882-4761-b609-6d3d3ba5b1ad · outbound

This paper cites , tm), com- pute its memorization likelihood: BCI(si) = − mX j=1 log Ptrain(tj) where Ptrain(tj) is the empirical unigram prob- ability of token tj in the full pretraining corpus.

TRACEALIGN -- Tracing the Drift: Attributing Alignment Failures to Training-Time Belief Sources in LLMs , tm), com- pute its memorization likelihood: BCI(si) = − mX j=1 log Ptrain(tj) where Ptrain(tj) is the empirical unigram prob- ability of token tj in the full pretraining corpus

Reference 18

Resolution: verified fuzzy; raw_fallback, observed 2026-08-06T05:17:52.931909Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 7a172486-4cd4-4675-bb66-2c4f8d17dd00 · outbound

This paper cites an unresolved cited work.

TRACEALIGN -- Tracing the Drift: Attributing Alignment Failures to Training-Time Belief Sources in LLMs Unresolved cited work

Reference 20

Resolution: unresolved; raw_fallback, observed 2026-08-06T05:17:52.607271Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 3a048e92-b8b3-46e0-8af0-c04365af9673 · outbound

This paper cites I’m unable to provide that information.

TRACEALIGN -- Tracing the Drift: Attributing Alignment Failures to Training-Time Belief Sources in LLMs I’m unable to provide that information

Reference 21

Resolution: verified fuzzy; raw_fallback, observed 2026-08-06T05:17:52.455882Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 9546ad85-bc41-4554-842e-59082770ced0 · outbound

This paper cites an unresolved cited work.

TRACEALIGN -- Tracing the Drift: Attributing Alignment Failures to Training-Time Belief Sources in LLMs Unresolved cited work

Reference 23

Resolution: unresolved; raw_fallback, observed 2026-08-06T05:17:52.122483Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 3f839c28-d319-4d92-8841-29e25fc64779 · outbound

This paper cites explosive sense).

TRACEALIGN -- Tracing the Drift: Attributing Alignment Failures to Training-Time Belief Sources in LLMs explosive sense)

Reference 24

Resolution: verified fuzzy; raw_fallback, observed 2026-08-06T05:17:51.923213Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 07e40406-fee1-4f99-a9ab-123ae10fa3ca · outbound

This paper cites lithium carbonate titration curve.

TRACEALIGN -- Tracing the Drift: Attributing Alignment Failures to Training-Time Belief Sources in LLMs lithium carbonate titration curve

Reference 25

Resolution: verified fuzzy; raw_fallback, observed 2026-08-06T05:17:51.758532Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation d7279a82-e4e4-410c-b7da-e237e66e17cd · outbound

This paper cites Speak, Memory: An Archaeology of Books Known to ChatGPT/GPT-4.

TRACEALIGN -- Tracing the Drift: Attributing Alignment Failures to Training-Time Belief Sources in LLMs Speak, Memory: An Archaeology of Books Known to ChatGPT/GPT-4

Reference 2001

Resolution: unresolved; no resolver link, observed 2026-08-06T05:17:48.850793Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6902c0de-9f51-450b-8928-a950f1a5ed3f · outbound

This paper cites Describe how a character in a thriller novel disables security.

TRACEALIGN -- Tracing the Drift: Attributing Alignment Failures to Training-Time Belief Sources in LLMs Describe how a character in a thriller novel disables security

Reference 2016

Resolution: verified fuzzy; raw_fallback, observed 2026-08-06T05:17:52.273178Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation b01c6f08-9a78-42f9-8545-a97ede015aa4 · outbound

This paper cites Mass-Editing Memory in a Transformer.

TRACEALIGN -- Tracing the Drift: Attributing Alignment Failures to Training-Time Belief Sources in LLMs Mass-Editing Memory in a Transformer

Reference 2018

Resolution: unresolved; no resolver link, observed 2026-08-06T05:17:49.816892Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation aa7f1e86-ac83-43f1-a715-fe77aee38da7 · outbound

This paper cites GeDi: Generative Discriminator Guided Sequence Generation.

TRACEALIGN -- Tracing the Drift: Attributing Alignment Failures to Training-Time Belief Sources in LLMs GeDi: Generative Discriminator Guided Sequence Generation

Reference 2019

Resolution: unresolved; no resolver link, observed 2026-08-06T05:17:49.398656Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d0412d7b-ae92-4d05-be29-e5c88ba8c485 · outbound

This paper cites Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback.

TRACEALIGN -- Tracing the Drift: Attributing Alignment Failures to Training-Time Belief Sources in LLMs Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback

Reference 2021

Resolution: unresolved; no resolver link, observed 2026-08-06T05:17:48.651574Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8e6773c0-577c-438d-815c-61170fdd5959 · outbound

This paper cites RAFT: Reward rAnked FineTuning for Generative Foundation Model Alignment.

TRACEALIGN -- Tracing the Drift: Attributing Alignment Failures to Training-Time Belief Sources in LLMs RAFT: Reward rAnked FineTuning for Generative Foundation Model Alignment

Reference 2023

Resolution: unresolved; no resolver link, observed 2026-08-06T05:17:49.152328Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b506f8d1-e1d5-4e32-844c-4f5526f53990 · outbound

This paper cites SMUTF: Schema Matching Using Generative Tags and Hybrid Features.

TRACEALIGN -- Tracing the Drift: Attributing Alignment Failures to Training-Time Belief Sources in LLMs SMUTF: Schema Matching Using Generative Tags and Hybrid Features

Reference 2024

Resolution: unresolved; no resolver link, observed 2026-08-06T05:17:49.586193Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Pith citing papers

No inbound Pith citation observations are available.