Paper Citation Record · LEDGER

How do language models learn facts? Dynamics, curricula and hallucinations

As of 22 August 2026, this Paper Citation Record lists 0 of 0 outbound references and 16 inbound Pith citation observations for arXiv:2503.21676.

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

pith.paper-citation-record.v1
2503.21676 v2

Coverage vector

measured 0 of 0 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links

measured 16 of 16 standing notices

One-hop event checks from named stored sources.

Source: scholarly_work_events, retraction_status_cache, observed 2026-08-22T06:32:14.747728+00:00

measured 16 of 16 inbound itemization

Pith citing papers itemized under the disclosed page cap.

Source: paper_references, paper_reference_links, observed 2026-08-15T21:55:30.393560Z

measured 0 of 1 external citation measurements

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

Source: arxiv_reference, observed 2026-07-04T21:10:09.137081Z

Reference resolution

0 of 0 outbound references displayed

verified exact0
verified fuzzy0
unresolved0
parse uncertain0
malformed identifier0
metadata mismatch0

External citation measurements

No source-named external measurement is stored.

Outbound references

No outbound reference observations are available for this paper version.

Pith citing papers

Observation 8b85d7fd-90ed-4166-a1c3-c4b06af2131d · inbound

Probability Consistency in Large Language Models: Theoretical Foundations Meet Empirical Discrepancies cites this paper.

Probability Consistency in Large Language Models: Theoretical Foundations Meet Empirical Discrepancies How do language models learn facts? Dynamics, curricula and hallucinations

Reference 38

Resolution: unresolved; no resolver link, observed 2026-08-15T21:55:30.393560Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 355ac051-c3fe-4392-9c0d-5218e9174566 · inbound

Pierce the Mists, Greet the Sky: Decipher Knowledge Overshadowing via Knowledge Circuit Analysis cites this paper.

Pierce the Mists, Greet the Sky: Decipher Knowledge Overshadowing via Knowledge Circuit Analysis How do language models learn facts? Dynamics, curricula and hallucinations

Reference 70

Resolution: unresolved; no resolver link, observed 2026-08-07T15:39:22.241145Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4c53945a-ff0b-4e66-9402-b5699aeb5601 · inbound

Do Activation Verbalization Methods Convey Privileged Information? cites this paper.

Do Activation Verbalization Methods Convey Privileged Information? How do language models learn facts? Dynamics, curricula and hallucinations

Reference 57

Resolution: metadata mismatch; arxiv_id, observed 2026-05-18T15:42:42.190199Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation ee5965d7-70ea-4a5d-93b3-42b8887e5c56 · inbound

How Training Data Shapes the Use of Parametric and In-Context Knowledge in Language Models cites this paper.

How Training Data Shapes the Use of Parametric and In-Context Knowledge in Language Models How do language models learn facts? Dynamics, curricula and hallucinations

Reference 20

Resolution: metadata mismatch; arxiv_id, observed 2026-05-18T13:11:23.889433Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 82e1efa7-3685-45cc-8181-57e0c18cc4c4 · inbound

Deep sequence models tend to memorize geometrically; it is unclear why cites this paper.

Deep sequence models tend to memorize geometrically; it is unclear why How do language models learn facts? Dynamics, curricula and hallucinations

Reference 211

Resolution: verified exact; arxiv_id, observed 2026-05-21T20:40:36.331364Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 60c5adf9-70ae-4f6d-b0da-a9645b3bba54 · inbound

Cram Less to Fit More: Training Data Pruning Improves Memorization of Facts cites this paper.

Cram Less to Fit More: Training Data Pruning Improves Memorization of Facts How do language models learn facts? Dynamics, curricula and hallucinations

Reference 103

Resolution: metadata mismatch; arxiv_id, observed 2026-05-11T06:15:59.034335Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 497a2c06-26de-4241-8be3-48459e55ba2d · inbound

Why Fine-Tuning Encourages Hallucinations and How to Fix It cites this paper.

Why Fine-Tuning Encourages Hallucinations and How to Fix It How do language models learn facts? Dynamics, curricula and hallucinations

Reference 4

Resolution: metadata mismatch; arxiv_id, observed 2026-05-10T10:55:04.180489Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 59052abd-1647-4cc8-95c7-8bfd79ae85be · inbound

Fine-Tuning Without Forgetting via Loss-Adaptive Learning Rates cites this paper.

Fine-Tuning Without Forgetting via Loss-Adaptive Learning Rates How do language models learn facts? Dynamics, curricula and hallucinations

Reference 76

Resolution: metadata mismatch; arxiv_id, observed 2026-05-20T07:18:07.006758Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 5b3d3ac5-b0c0-4b92-8ccf-eab1a68a0079 · inbound

The Future of Facts: Tracing the Factual Generation-Verification Gap cites this paper.

The Future of Facts: Tracing the Factual Generation-Verification Gap How do language models learn facts? Dynamics, curricula and hallucinations

Reference 33

Resolution: metadata mismatch; arxiv_id, observed 2026-06-29T18:33:50.386625Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation dd67d6de-15a9-46ea-9a9c-f0ae3d96e94b · inbound

Natural Ungrokking: Asymmetric Control of Which Rules Survive Pretraining cites this paper.

Natural Ungrokking: Asymmetric Control of Which Rules Survive Pretraining How do language models learn facts? Dynamics, curricula and hallucinations

Reference 48

Resolution: metadata mismatch; arxiv_id, observed 2026-07-04T21:10:09.138640Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 439c4b36-fc36-47ab-94af-407f733aac6b · inbound

Structure Before Collapse: Transient semantic geometry in next-token prediction cites this paper.

Structure Before Collapse: Transient semantic geometry in next-token prediction How do language models learn facts? Dynamics, curricula and hallucinations

Reference 14

Resolution: metadata mismatch; arxiv_id, observed 2026-07-04T13:29:51.028707Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-22T06:32:14.747728+00:00.

Observation 05131139-e569-4c02-84bd-af6376d4ffa5 · inbound

Pretraining Curricula Enable Selective Fine-tuning cites this paper.

Pretraining Curricula Enable Selective Fine-tuning How do language models learn facts? Dynamics, curricula and hallucinations

Reference 76

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 71e586a5-5b8d-421f-88f1-3647ebbb8448 · inbound

Can a Language Model Learn Facts Continually in Its Weights? cites this paper.

Can a Language Model Learn Facts Continually in Its Weights? How do language models learn facts? Dynamics, curricula and hallucinations

Reference 25

Resolution: unresolved; no resolver link, observed 2026-07-14T07:36:43.499258Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e0beb664-6de5-43d1-83e9-547d15ee94a4 · inbound

Reliability Scales Inversely: Hallucinations Snowball Faster in Bigger Language Models cites this paper.

Reliability Scales Inversely: Hallucinations Snowball Faster in Bigger Language Models How do language models learn facts? Dynamics, curricula and hallucinations

Reference 46

Resolution: unresolved; no resolver link, observed 2026-08-02T09:19:44.855997Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9e794fc0-d4b3-4f3e-8689-57363cf99c71 · inbound

Bridging Compute- and Data-Optimal Pretraining cites this paper.

Bridging Compute- and Data-Optimal Pretraining How do language models learn facts? Dynamics, curricula and hallucinations

Reference 72

Resolution: unresolved; no resolver link, observed 2026-08-01T03:02:04.276215Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 52e367e9-605e-402b-bcd2-e0bf1a675f20 · inbound

Training AI Scientists to Replicate Research cites this paper.

Training AI Scientists to Replicate Research How do language models learn facts? Dynamics, curricula and hallucinations

Reference 5

Resolution: unresolved; no resolver link, observed 2026-08-14T13:35:09.490618Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.