Pith. sign in

Paper Citation Record · LEDGER

Through the Bottleneck: How Multi-head Latent Attention Separates Content from Position in Language Models

As of 9 August 2026, this Paper Citation Record lists 21 of 21 outbound references and 0 inbound Pith citation observations for arXiv:2607.23054.

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

pith.paper-citation-record.v1
2607.23054 v1

Coverage vector

measured 21 of 21 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-08-01T03:47:08.170882Z

measured 21 of 21 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

21 of 21 outbound references displayed

  • verified exact0
  • verified fuzzy0
  • unresolved21
  • parse uncertain0
  • malformed identifier0
  • metadata mismatch0

External citation measurements

No source-named external measurement is stored.

Outbound references

Observation 04a3af1f-7721-4121-abe6-810008269a09 · outbound

This paper cites Ainslie, J.

Through the Bottleneck: How Multi-head Latent Attention Separates Content from Position in Language Models Ainslie, J

Reference 1

Resolution
unresolved
no resolver link, observed 2026-08-01T03:47:06.300071Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=arxiv_source observed=2026-08-01T03:47:06.300071Z digest=sha256:549a08a7c80f2677f0087ad34a126d886df5d6fd7b53d73c4878e70b389f3148

Observation 77afe194-3328-4450-8a85-e5e49188a18d · outbound

This paper cites Alain and Y.

Through the Bottleneck: How Multi-head Latent Attention Separates Content from Position in Language Models Alain and Y

Reference 2

Resolution
unresolved
no resolver link, observed 2026-08-01T03:47:06.375962Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=arxiv_source observed=2026-08-01T03:47:06.375962Z digest=sha256:e26cecc147df3400ce98543c2b7cac7ddbe71e5da9685c50de59251796636345

Observation bf903957-fa06-470e-ae9f-9b1984ff902a · outbound

This paper cites Belinkov, N.

Through the Bottleneck: How Multi-head Latent Attention Separates Content from Position in Language Models Belinkov, N

Reference 3

Resolution
unresolved
no resolver link, observed 2026-08-01T03:47:06.450986Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=arxiv_source observed=2026-08-01T03:47:06.450986Z digest=sha256:0d5df63a4b862d5448dc9b559a7b6abc69c06b7a33e2b1168139bd55419ba642

Observation d260c136-a24a-4c2e-8161-6003765d1e73 · outbound

This paper cites Bricken, A.

Through the Bottleneck: How Multi-head Latent Attention Separates Content from Position in Language Models Bricken, A

Reference 4

Resolution
unresolved
no resolver link, observed 2026-08-01T03:47:06.536730Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=arxiv_source observed=2026-08-01T03:47:06.536730Z digest=sha256:e12b4c169f42ba03a50a92cf03b502498a1e588aacf2fa6e9b760a949fc7c399

Observation 627d47c1-0cdc-4212-b8b6-4c0d323e0dea · outbound

This paper cites Clark, U.

Through the Bottleneck: How Multi-head Latent Attention Separates Content from Position in Language Models Clark, U

Reference 5

Resolution
unresolved
no resolver link, observed 2026-08-01T03:47:06.740846Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=arxiv_source observed=2026-08-01T03:47:06.740846Z digest=sha256:01104b08291a07378e61c7e2cd269d11171f1f4b19ea1f908029f71e9fd0b97b

Observation c5661f29-3dc7-4dfc-b32d-453f8acaba9a · outbound

This paper cites Conmy, A.

Through the Bottleneck: How Multi-head Latent Attention Separates Content from Position in Language Models Conmy, A

Reference 6

Resolution
unresolved
no resolver link, observed 2026-08-01T03:47:06.903779Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=arxiv_source observed=2026-08-01T03:47:06.903779Z digest=sha256:176f0a8b8480a5f6496abd547bc77124e488e977bd81506b6c4cb16860caba94

Observation bec3a37b-6c65-48df-a151-8d9074d91573 · outbound

This paper cites Cunningham, A.

Through the Bottleneck: How Multi-head Latent Attention Separates Content from Position in Language Models Cunningham, A

Reference 7

Resolution
unresolved
no resolver link, observed 2026-08-01T03:47:06.951932Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=arxiv_source observed=2026-08-01T03:47:06.951932Z digest=sha256:17b7ef62e879400a8a774a5ec49d761a6c2c6407e9b7ecd680538a4fdcdf5498

Observation 1b984e2f-25c6-4657-9d45-a31f75986b8c · outbound

This paper cites DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model.

Through the Bottleneck: How Multi-head Latent Attention Separates Content from Position in Language Models DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model

Reference 8

Resolution
unresolved
no resolver link, observed 2026-08-01T03:47:06.996718Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=arxiv_source observed=2026-08-01T03:47:06.996718Z digest=sha256:4b8b05756a6472bbaf1cea54ec71f002f804dbb024d6ac74954ef9c5f09005da

Observation 063a0297-e8bd-4c78-a601-63fefd94a17b · outbound

This paper cites DeepSeek-V3 Technical Report.

Through the Bottleneck: How Multi-head Latent Attention Separates Content from Position in Language Models DeepSeek-V3 Technical Report

Reference 9

Resolution
unresolved
no resolver link, observed 2026-08-01T03:47:07.104250Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=arxiv_source observed=2026-08-01T03:47:07.104250Z digest=sha256:ee529fcd72da3a0fe0f39eec5fa38e6e52e7319c9b091c47271c9d4b6a05c15b

Observation 840f7dbb-c781-419e-b7d7-effc2702d31d · outbound

This paper cites Eldan and Y.

Through the Bottleneck: How Multi-head Latent Attention Separates Content from Position in Language Models Eldan and Y

Reference 10

Resolution
unresolved
no resolver link, observed 2026-08-01T03:47:07.196846Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=arxiv_source observed=2026-08-01T03:47:07.196846Z digest=sha256:bd448bfe1e7cc8203d9ae475bd08f3159c29cff85f7ed6a95994c5a282d379e8

Observation 1751a1d7-802f-47d3-895a-c00550814b3b · outbound

This paper cites Elhage, N.

Through the Bottleneck: How Multi-head Latent Attention Separates Content from Position in Language Models Elhage, N

Reference 11

Resolution
unresolved
no resolver link, observed 2026-08-01T03:47:07.349832Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=arxiv_source observed=2026-08-01T03:47:07.349832Z digest=sha256:3610865a282a6817147bb2074842d795274fb9af3c8941ccc0efb395cfa14257

Observation f4eefd06-7d6f-4b08-90e8-6550b0b04102 · outbound

This paper cites Elhage, T.

Through the Bottleneck: How Multi-head Latent Attention Separates Content from Position in Language Models Elhage, T

Reference 12

Resolution
unresolved
no resolver link, observed 2026-08-01T03:47:07.460337Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=arxiv_source observed=2026-08-01T03:47:07.460337Z digest=sha256:6a66770111222e97f5241c47217c339d8644d585360f8e75ced1a0635104fcf3

Observation da7a03ce-50d7-44bc-8722-0e56b85b514d · outbound

This paper cites Hewitt and C.

Through the Bottleneck: How Multi-head Latent Attention Separates Content from Position in Language Models Hewitt and C

Reference 13

Resolution
unresolved
no resolver link, observed 2026-08-01T03:47:07.557965Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=arxiv_source observed=2026-08-01T03:47:07.557965Z digest=sha256:5b9fa7db9348a495278300bb54e197ceeb10d98ae1cdf474792a42608d486127

Observation 8b072cb2-7ddc-4f7e-969d-a1058fa8e5c6 · outbound

This paper cites The Stack: 3 TB of permissively licensed source code.

Through the Bottleneck: How Multi-head Latent Attention Separates Content from Position in Language Models The Stack: 3 TB of permissively licensed source code

Reference 14

Resolution
unresolved
no resolver link, observed 2026-08-01T03:47:07.597461Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=arxiv_source observed=2026-08-01T03:47:07.597461Z digest=sha256:9127f612c58d28c7b0a9e3d006640e1283532764063a60d086b951836f5380b7

Observation 2fc34cf0-85c8-4957-9809-41bb8c9611dd · outbound

This paper cites an unresolved cited work.

Through the Bottleneck: How Multi-head Latent Attention Separates Content from Position in Language Models Unresolved cited work

Reference 15

Resolution
unresolved
no resolver link, observed 2026-08-01T03:47:07.694248Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=arxiv_source observed=2026-08-01T03:47:07.694248Z digest=sha256:a75ba96a38f08f76dd07f27f7e5229277eeb4fba848ffc21cdac752386194a96

Observation 787f0dba-7d3c-4322-80c2-d99414f42298 · outbound

This paper cites Nanda, L.

Through the Bottleneck: How Multi-head Latent Attention Separates Content from Position in Language Models Nanda, L

Reference 16

Resolution
unresolved
no resolver link, observed 2026-08-01T03:47:07.761761Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=arxiv_source observed=2026-08-01T03:47:07.761761Z digest=sha256:097d1bb2ae0b930202f14a07b2ec378f7a99d74bb6d8a3bc96bf74940912c39a

Observation e2e62617-879a-4834-b7e3-a0bee4e81565 · outbound

This paper cites Olsson, N.

Through the Bottleneck: How Multi-head Latent Attention Separates Content from Position in Language Models Olsson, N

Reference 17

Resolution
unresolved
no resolver link, observed 2026-08-01T03:47:07.830220Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=arxiv_source observed=2026-08-01T03:47:07.830220Z digest=sha256:1b1c6f0d732503a6ea701de961516a6d24e9c86ae0ce01272d92d387b8d7cd90

Observation 65b5b572-436f-4b4d-9fa1-b4c2880fbf4d · outbound

This paper cites OpenWebMath: An Open Dataset of High-Quality Mathematical Web Text.

Through the Bottleneck: How Multi-head Latent Attention Separates Content from Position in Language Models OpenWebMath: An Open Dataset of High-Quality Mathematical Web Text

Reference 18

Resolution
unresolved
no resolver link, observed 2026-08-01T03:47:07.931681Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=arxiv_source observed=2026-08-01T03:47:07.931681Z digest=sha256:e2a7ece94b96de399b6133cb9011ed854ea87eb9859a15cc0443dbf2976d3e5b

Observation f939ac05-98b0-4211-b3d4-f00bd6b4390f · outbound

This paper cites The FineWeb Datasets: Decanting the Web for the Finest Text Data at Scale.

Through the Bottleneck: How Multi-head Latent Attention Separates Content from Position in Language Models The FineWeb Datasets: Decanting the Web for the Finest Text Data at Scale

Reference 19

Resolution
unresolved
no resolver link, observed 2026-08-01T03:47:08.032030Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=arxiv_source observed=2026-08-01T03:47:08.032030Z digest=sha256:c84906f77907199a404d5f782c55c638566caed700e24c6cad69fbf4ca158729

Observation dbe50eee-48eb-4fa5-96f5-7cbbcfd4f758 · outbound

This paper cites Fast Transformer Decoding: One Write-Head is All You Need.

Through the Bottleneck: How Multi-head Latent Attention Separates Content from Position in Language Models Fast Transformer Decoding: One Write-Head is All You Need

Reference 20

Resolution
unresolved
no resolver link, observed 2026-08-01T03:47:08.098841Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=arxiv_source observed=2026-08-01T03:47:08.098841Z digest=sha256:4d7f1865a179d6cde606882c70382bf5c6efd5f9670efa158a34d88d0f022559

Observation 6848c6d8-bf6f-440e-b3e8-017182fc383c · outbound

This paper cites an unresolved cited work.

Through the Bottleneck: How Multi-head Latent Attention Separates Content from Position in Language Models Unresolved cited work

Reference 21

Resolution
unresolved
no resolver link, observed 2026-08-01T03:47:08.170882Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=arxiv_source observed=2026-08-01T03:47:08.170882Z digest=sha256:3ae10708ee4ea9a53112b4d8dd5313d9bcf3b19d091d80c25a9bfffd5688a72c

Pith citing papers

No inbound Pith citation observations are available.