Pith. sign in

Paper Citation Record · LEDGER

How Bad is Training on Synthetic Data? A Statistical Analysis of Language Model Collapse

As of 7 August 2026, this Paper Citation Record lists 0 of 0 outbound references and 14 inbound Pith citation observations for arXiv:2404.05090.

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

pith.paper-citation-record.v1
2404.05090 v1

Coverage vector

measured 0 of 0 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links

measured 14 of 14 standing notices

One-hop event checks from named stored sources.

Source: scholarly_work_events, retraction_status_cache, observed 2026-08-07T06:34:17.273281+00:00

measured 14 of 14 inbound itemization

Pith citing papers itemized under the disclosed page cap.

Source: paper_references, paper_reference_links, observed 2026-08-07T13:59:48.929912Z

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

3
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 41537235-2c53-4a97-a0a3-da565367e547 · inbound

Using Sign Language Production as Data Augmentation to enhance Sign Language Translation cites this paper.

Using Sign Language Production as Data Augmentation to enhance Sign Language Translation How Bad is Training on Synthetic Data? A Statistical Analysis of Language Model Collapse

Reference 66

Resolution
unresolved
no resolver link, observed 2026-08-07T04:47:50.676984Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-07T04:47:50.676984Z digest=sha256:1df71e80b03355906bfa130a59409043dabfa6c9ad4b7bdc9aabbca353050795

Observation da71ca6e-0197-49cb-a9c3-d7a7f5be4859 · inbound

LLM Web Dynamics: Tracing Model Collapse in a Network of LLMs cites this paper.

LLM Web Dynamics: Tracing Model Collapse in a Network of LLMs How Bad is Training on Synthetic Data? A Statistical Analysis of Language Model Collapse

Reference 33

Resolution
unresolved
no resolver link, observed 2026-08-07T13:59:48.929912Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=arxiv_source observed=2026-08-07T13:59:48.929912Z digest=sha256:199be65bf7e0c1922f0e60af813730fd3eec18bcb91287fba24ce9e3f8102536

Observation 9e77a416-a27a-4afa-88bf-749bc2797c13 · inbound

What Matters in LLM-generated Data: Diversity and Its Effect on Model Fine-Tuning cites this paper.

What Matters in LLM-generated Data: Diversity and Its Effect on Model Fine-Tuning How Bad is Training on Synthetic Data? A Statistical Analysis of Language Model Collapse

Reference 30

Resolution
unresolved
no resolver link, observed 2026-08-06T23:12:18.999831Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-06T23:12:18.999831Z digest=sha256:89d2f7ce7c3e4486fa03af5c22dd8188ab187168e4442ad21d2e7f4628e5c187

Observation e9860435-a193-4322-89ae-96683773c6f5 · inbound

Unlocking Speech Instruction Data Potential with Query Rewriting cites this paper.

Unlocking Speech Instruction Data Potential with Query Rewriting How Bad is Training on Synthetic Data? A Statistical Analysis of Language Model Collapse

Reference 40

Resolution
unresolved
no resolver link, observed 2026-08-06T18:21:34.589888Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=arxiv_source observed=2026-08-06T18:21:34.589888Z digest=sha256:81758c78c49f57e595b53999dc29058aa874060a760a7ea7bf8d7cbd63460b77

Observation de233454-e05b-4173-837b-d657cdc9e7b2 · inbound

A Penalty Goes a Long Way: Measuring Lexical Diversity in Synthetic Texts Under Prompt-Influenced Length Variations cites this paper.

A Penalty Goes a Long Way: Measuring Lexical Diversity in Synthetic Texts Under Prompt-Influenced Length Variations How Bad is Training on Synthetic Data? A Statistical Analysis of Language Model Collapse

Reference 32

Resolution
unresolved
no resolver link, observed 2026-08-06T15:46:27.825071Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=arxiv_source observed=2026-08-06T15:46:27.825071Z digest=sha256:50b9537ad42a347af3323fc96cb8e6abc1a620b0043d7ab4ceaab6667d8f5b69

Observation 8bd8742f-9ea8-4afa-882f-10fa73018190 · inbound

A Generative Foundation Model for Chest Radiography cites this paper.

A Generative Foundation Model for Chest Radiography How Bad is Training on Synthetic Data? A Statistical Analysis of Language Model Collapse

Reference 33

Resolution
unresolved
no resolver link, observed 2026-08-05T10:38:07.338196Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-05T10:38:07.338196Z digest=sha256:c5607d576bcb423e0e0c4bc1d304f81157f12148a875ffd30db601fc2a20ce74

Observation 4f6ba3b3-58aa-49b0-ad13-64481b3a74e9 · inbound

The Anti-Ouroboros Effect: Emergent Resilience in Large Language Models from Recursive Selective Feedback cites this paper.

The Anti-Ouroboros Effect: Emergent Resilience in Large Language Models from Recursive Selective Feedback How Bad is Training on Synthetic Data? A Statistical Analysis of Language Model Collapse

Reference 4

Resolution
unresolved
no resolver link, observed 2026-08-05T12:04:03.464968Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-05T12:04:03.464968Z digest=sha256:46b9380977e32bcd55a32affd542ca53ebbedb8e2a4acecbf37a17d6fcfb9dc6

Observation d301647f-202c-40ee-812f-fd91324aeb54 · inbound

Generative artificial intelligence reduces social welfare through model collapse cites this paper.

Generative artificial intelligence reduces social welfare through model collapse How Bad is Training on Synthetic Data? A Statistical Analysis of Language Model Collapse

Reference 36

Resolution
verified exact
arxiv_id, observed 2026-05-11T18:56:05.954915Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.

source=pdf_text observed=2026-05-08T13:26:57.925635Z digest=sha256:5c0e822ee3b2e7d9bf221e14f4dfdd602f5d3a4cd269b126ee917613128faabd

Observation e8c76f88-3021-4b45-bfa5-d9d524545786 · inbound

Iterative Finetuning is Mostly Idempotent cites this paper.

Iterative Finetuning is Mostly Idempotent How Bad is Training on Synthetic Data? A Statistical Analysis of Language Model Collapse

Reference 9

Resolution
metadata mismatch
arxiv_id, observed 2026-05-11T15:51:44.493288Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.

source=pdf_text observed=2026-05-09T19:04:36.210200Z digest=sha256:83b172f118c323890add91703d24837236775c322f713d3a5f72df87b5405f64

Observation 8a7e45be-ce65-4361-8a97-09e0b8041153 · inbound

Position: the Stochastic Parrot in the Coal Mine. Model Collapse is a Threat to Low-Resource Communities cites this paper.

Position: the Stochastic Parrot in the Coal Mine. Model Collapse is a Threat to Low-Resource Communities How Bad is Training on Synthetic Data? A Statistical Analysis of Language Model Collapse

Reference 28

Resolution
verified exact
arxiv_id, observed 2026-05-09T06:35:40.856430Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.

source=pdf_text observed=2026-05-08T18:16:51.881163Z digest=sha256:b9d1d30232fe460dceeaaf4f532543c23244292739aca790c754a93a6441bcee

Observation a1682d1f-71b2-41c1-8e89-604c1c312f94 · inbound

Self-Training Doesn't Flatten Language -- It Restructures It: Surface Markers Amplify While Deep Syntax Dies cites this paper.

Self-Training Doesn't Flatten Language -- It Restructures It: Surface Markers Amplify While Deep Syntax Dies How Bad is Training on Synthetic Data? A Statistical Analysis of Language Model Collapse

Reference 22

Resolution
verified exact
arxiv_id, observed 2026-05-21T05:49:41.082711Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.

source=arxiv_source observed=2026-05-21T05:46:07.176408Z digest=sha256:d7ce2ee90ae8eda4b1ea30cf70da9d3af2f9ed1e4d070ec049e94f099ea54aeb

Observation bd86a422-b354-42b2-a0d1-f1764d53e3c9 · inbound

Model Collapse as Cultural Evolution cites this paper.

Model Collapse as Cultural Evolution How Bad is Training on Synthetic Data? A Statistical Analysis of Language Model Collapse

Reference 43

Resolution
verified exact
arxiv_id, observed 2026-05-25T05:35:23.455429Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.

source=arxiv_source observed=2026-05-25T05:30:31.504863Z digest=sha256:78605c329edfe93aed31896daefd2ce9e4138d1c0bedcc05f6c52dc9d2eb5ce5

Observation 0741d3d5-8942-4747-b3ab-676ee7135479 · inbound

Activation Steering for Synthetic Data Generation: The Role of Diversity in Downstream Safety Detection cites this paper.

Activation Steering for Synthetic Data Generation: The Role of Diversity in Downstream Safety Detection How Bad is Training on Synthetic Data? A Statistical Analysis of Language Model Collapse

Reference 43

Resolution
verified exact
arxiv_id, observed 2026-06-29T14:03:29.933958Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.

source=pdf_text observed=2026-06-29T13:53:27.306664Z digest=sha256:d7a248db99d8fce392d09d247990ef5cc4cc8bf3b52194e6dac1e39f26ab88b9

Observation 7f84393b-0640-4913-883a-bf65095ac936 · inbound

When AI Reviews Its Own Code: Recursive Self-Training Collapse in Code LLMs cites this paper.

When AI Reviews Its Own Code: Recursive Self-Training Collapse in Code LLMs How Bad is Training on Synthetic Data? A Statistical Analysis of Language Model Collapse

Reference 151

Resolution
verified exact
arxiv_id, observed 2026-06-30T01:34:09.458265Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.

source=arxiv_source observed=2026-06-30T01:29:42.919461Z digest=sha256:16a8418494907ec34bdc5cae8f4cd61e5dfa8e2226006ccc5d995f95f7354499