Pith. sign in

Paper Citation Record · LEDGER

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers

As of 4 August 2026, this Paper Citation Record lists 51 of 51 outbound references and 0 inbound Pith citation observations for arXiv:2605.03780.

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

pith.paper-citation-record.v1
2605.03780 v1

Coverage vector

measured 51 of 51 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-05-07T16:37:26.420806Z

measured 51 of 51 standing notices

One-hop event checks from named stored sources.

Source: scholarly_work_events, retraction_status_cache, observed 2026-08-03T06:30:56.289259+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

51 of 51 outbound references displayed

  • verified exact11
  • verified fuzzy34
  • unresolved5
  • parse uncertain0
  • malformed identifier0
  • metadata mismatch1

External citation measurements

No source-named external measurement is stored.

Outbound references

Observation 5946865f-e167-4f49-a600-d0fee130acd2 · outbound

This paper cites Aitchison.The Statistical Analysis of Compositional Data.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Aitchison.The Statistical Analysis of Compositional Data

Reference 1

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.641748Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:0fa88370ce31348fc76cc72d10a27a01ab004e5740f92e5604e4834744f0724f

Observation e8264c18-1833-4a9c-a6dd-e30605ccb3b9 · outbound

This paper cites What learning algorithm is in-context learning? Investigations with linear models.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers What learning algorithm is in-context learning? Investigations with linear models

Reference 2

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.658225Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:3d0a465f177bb9867538a67c558c568ec7a3bf05ad277b3d86bf652f26cdeb39

Observation b20da21c-77da-462c-8621-de9d99fc4dd6 · outbound

This paper cites Towards monosemanticity: Decomposing language models with dictionary learning.Transformer Circuits Thread.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Towards monosemanticity: Decomposing language models with dictionary learning.Transformer Circuits Thread

Reference 3

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.662383Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:447fff17bada70f6f9935ddebb067d0af1d0581bdde45a7ed723d1400d1aa732

Observation bda420c7-81ac-4997-9280-42ed0aed7bdd · outbound

This paper cites Language models are few-shot learners.Advances in neural information processing systems, 33:1877–1901.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Language models are few-shot learners.Advances in neural information processing systems, 33:1877–1901

Reference 4

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.634026Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:fe1db225b11b682353e1b252cab26b010ab6571b4a639872cea682f141a13168

Observation 17df5d75-c7a7-442b-a989-c4916bc2e55e · outbound

This paper cites Robust uncertainty principles: Exact signal reconstruction from highly incomplete frequency information.IEEE Transactions on information theory, 52(2):489–509.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Robust uncertainty principles: Exact signal reconstruction from highly incomplete frequency information.IEEE Transactions on information theory, 52(2):489–509

Reference 5

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.629724Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:d41ee8b3bbed1c5bf9739bb5ce6959721ca78c1a3d40f7a98f7a872e599082ac

Observation b1a2742b-4dcc-4a52-84ff-c41e2239652d · outbound

This paper cites Dynamics of Transient Structure in In-Context Linear Regression Transformers.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Dynamics of Transient Structure in In-Context Linear Regression Transformers

Reference 6

Resolution
verified exact
arxiv_id, observed 2026-05-11T23:36:36.351014Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:9f3d48c8f3d34795aec700c750546f494433cdc1f3ad1a58a83f7ee0e5bc21f5

Observation 6de3b744-2f48-489a-b346-82906294c5c4 · outbound

This paper cites Compressed sensing.IEEE Transactions on information theory, 52(4): 1289–1306.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Compressed sensing.IEEE Transactions on information theory, 52(4): 1289–1306

Reference 7

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.625138Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:0a2a8f0d4c514899e47a220aca1d62bf22e94afdabe6aecc34d2a5a5b2976a05

Observation 892b958c-1366-4c55-8548-3b36de5fbd35 · outbound

This paper cites Uncertainty principles and signal recovery.SIAM Journal on Applied Mathematics, 49(3):906–931.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Uncertainty principles and signal recovery.SIAM Journal on Applied Mathematics, 49(3):906–931

Reference 8

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.648176Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:21f81521dc986bcd4fa7135075d112cf626a7e4db8d82dad1f0a48e5a023ceef

Observation f9222229-9c05-4182-a7d3-7475c9e7f834 · outbound

This paper cites Duchi, Shai Shalev-Shwartz, Yoram Singer, and Tushar Chandra.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Duchi, Shai Shalev-Shwartz, Yoram Singer, and Tushar Chandra

Reference 9

Resolution
verified exact
arxiv_id, observed 2026-05-09T01:54:34.595706Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:0e73d66ee06d37c4ebaf4c0b04f27e8bc584c82d1c0347ceab73b5ee342dfe58

Observation 1e93c86b-ca72-4d10-8883-3f5ff4da4184 · outbound

This paper cites The evolution of statistical induction heads: In-context learning markov chains.Advances in neural information processing systems, 37:64273–64311.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers The evolution of statistical induction heads: In-context learning markov chains.Advances in neural information processing systems, 37:64273–64311

Reference 10

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.654475Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:2ea80d5be566bff2a838a385d9048849ba9100d47dcf02053deb3643df5ef698

Observation 2aaa826f-5fde-4003-ba9c-add5d6408454 · outbound

This paper cites A mathematical framework for transformer circuits.Transformer Circuits Thread.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers A mathematical framework for transformer circuits.Transformer Circuits Thread

Reference 11

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.638033Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:9a743ccfb47e3b10936c5cd6f6d63182edfdd8d150742bd99bbdecc9da686702

Observation 8d145da7-130f-42e3-bb43-13a778881138 · outbound

This paper cites an unresolved cited work.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Unresolved cited work

Reference 12

Resolution
unresolved
raw_fallback, observed 2026-05-27T02:03:23.644817Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:2de42ef51b5a39f953f73c92c495993bc0cccb902f91312e93642d2423088d8c

Observation f8be15f7-67bd-48fb-aa8e-b4fef03cffb8 · outbound

This paper cites Toy models of superposition.Transformer Circuits Thread.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Toy models of superposition.Transformer Circuits Thread

Reference 13

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.559586Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:e8ea40cafa900f4d687cae994ecaccc6c6b066dae8f6847b6ac007555bcde030

Observation 07140705-bf44-4525-8540-770fc5496d3a · outbound

This paper cites What can transformers learn in-context? a case study of simple function classes.Advances in Neural Information Processing Systems, 35:30583–30598.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers What can transformers learn in-context? a case study of simple function classes.Advances in Neural Information Processing Systems, 35:30583–30598

Reference 14

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.553274Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:767e231b743972c0789a59f3ec080ad2222c825708e9d2cde2292f27d4738772

Observation 32cec3f9-cc17-4526-9886-9a37055ef4f0 · outbound

This paper cites Transformer feed-forward layers are key-value memories.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Transformer feed-forward layers are key-value memories

Reference 15

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.563789Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:a6ae8d19f94886faf0efd15f195ca19f4423ca9e4b072df72dc396238a777416

Observation 6aef1235-57f5-41bf-95a3-0a9f25d3a921 · outbound

This paper cites When models manipulate manifolds: The geometry of a counting task.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers When models manipulate manifolds: The geometry of a counting task

Reference 16

Resolution
verified exact
arxiv_id, observed 2026-05-11T23:36:36.360333Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:adf98ad3b5556b3c5a39457693a82f9522fcff5fefb40c8f7cbd3a73b44dd81c

Observation 688e331c-3bc1-4bc6-932c-b44eec69526f · outbound

This paper cites In-context learning creates task vectors.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers In-context learning creates task vectors

Reference 17

Resolution
verified exact
doi, observed 2026-05-09T01:54:34.562396Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:9ec7a4190631d5f38aa1dbf9483833cd0a83c614a169e03d6a79ceae37cad27f

Observation ddd6fc35-b877-4fdc-8508-1f9a75fb6080 · outbound

This paper cites In-context learning state vector with inner and momentum optimiza- tion.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers In-context learning state vector with inner and momentum optimiza- tion

Reference 18

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.578115Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:e4417ba9927548952a4785675580cb17efa56b7d315581dbbb4b8d7181c3fa2a

Observation 6ef1d277-04af-42f1-a88f-179c32240473 · outbound

This paper cites an unresolved cited work.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Unresolved cited work

Reference 19

Resolution
unresolved
raw_fallback, observed 2026-05-27T02:03:23.581117Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:978d9104ac2f2201e9f0abd44ffb900268b38a88ea8be51beab807ee84f48175

Observation ae2f3838-cc85-4569-88bb-6c3e376be33c · outbound

This paper cites an unresolved cited work.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Unresolved cited work

Reference 20

Resolution
unresolved
raw_fallback, observed 2026-05-27T02:03:23.537073Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:2d1f8ecda36e13ac026e60637d227568a658701e091eac933f3a1d8403ea3969

Observation 8ed046fc-6e36-460c-a363-24ad344c3f6d · outbound

This paper cites Decoupled weight decay regularization.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Decoupled weight decay regularization

Reference 21

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.603829Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:6441e10e6be68c6187d363db80b32c79cebf530182d8800a3442f24b450da2b3

Observation 71c3ce90-d393-4d38-9060-746dc57cb336 · outbound

This paper cites Asymp- totic theory of in-context learning by linear attention.Proceedings of the National Academy of Sciences, 122(28):e2502599122.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Asymp- totic theory of in-context learning by linear attention.Proceedings of the National Academy of Sciences, 122(28):e2502599122

Reference 22

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.527272Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:b78f0986090b7e7b1bae93012cccc2200939713cb845e5a6edaf8c5cecf9bbbd

Observation 586704c3-d760-47fe-9e12-e8685077edf6 · outbound

This paper cites Language Models Implement Simple W ord2 V ec-style Vector Arithmetic.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Language Models Implement Simple W ord2 V ec-style Vector Arithmetic

Reference 23

Resolution
verified exact
doi, observed 2026-05-09T01:54:34.549977Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:401340648e2badf325e9b0819669bc45825a50185eadc55bdd375516ceb98555

Observation 5f9d748b-cd26-4b7a-9834-63ff087195cd · outbound

This paper cites Distributed representations of words and phrases and their compositionality.Advances in neural information processing systems, 26.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Distributed representations of words and phrases and their compositionality.Advances in neural information processing systems, 26

Reference 24

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.567668Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:88d5a6ae2af70f789f0cf826ccdba29d553e5b383e65ac311ca87904c0e5d2a8

Observation a7c841d5-ce6c-4435-8faa-db757f4efca7 · outbound

This paper cites Linguistic regularities in continuous space word representations.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Linguistic regularities in continuous space word representations

Reference 25

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.543403Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:7dd553ee3b42d2160f03d6a6a96b822829c72a5f383e5ae479b15e12ca42378f

Observation 78d5a1b8-bd9c-47c4-ac3c-16b6254db9d4 · outbound

This paper cites In-context Learning and Induction Heads.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers In-context Learning and Induction Heads

Reference 26

Resolution
verified exact
local_arxiv, observed 2026-05-11T23:36:36.373738Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:6b6993edce7cf51bf6fc0f6984b5875b9a8d7d7213f17929eb54e410152c508a

Observation e2d07be3-5dfd-41a0-b42d-4b71212e8109 · outbound

This paper cites What in-context learning “learns” in-context: Disentangling task recognition and task learning.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers What in-context learning “learns” in-context: Disentangling task recognition and task learning

Reference 27

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.593182Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:386c3adc8e11f893d7d7cff8c8f847d17fe77a5cd5536323d8811eff68e3d759

Observation 84efb644-3587-42c5-9cc8-909b96378d56 · outbound

This paper cites Competition dynamics shape algorithmic phases of in-context learning.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Competition dynamics shape algorithmic phases of in-context learning

Reference 28

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.599964Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:8a3e7cc508ba2a15adebceec75785a232890805c350fc0b7b88aa0bfd59b1d43

Observation 33c5dc3e-51c7-4b6c-a651-7275d95e8319 · outbound

This paper cites The linear representation hypothesis and the geometry of large language models.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers The linear representation hypothesis and the geometry of large language models

Reference 29

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.556271Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:c39e85eff141517ddf8848b48f9520a5b9073918f394842b546308b224b6f6d7

Observation 43664998-302c-42ee-a830-693c0089aa9c · outbound

This paper cites The geometry of categorical and hierarchical concepts in large language models.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers The geometry of categorical and hierarchical concepts in large language models

Reference 30

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.596481Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:1837093ec12f43bd0d44cb53b458926714c62297c34d23b8b2f4bf63aa085960

Observation 4ef955ce-2ab7-4b43-b5bb-144ddd4c4e56 · outbound

This paper cites Approximation theory of the mlp model in neural networks.Acta numerica, 8: 143–195.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Approximation theory of the mlp model in neural networks.Acta numerica, 8: 143–195

Reference 31

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.584470Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:f4ea6fcfec572dda6ec77695f12d607b75e6c03c9c5d8ee8f5ae2bfd0e317d82

Observation 49232a37-960f-4b69-ad67-8089ee24791f · outbound

This paper cites Qwen2.5 Technical Report.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Qwen2.5 Technical Report

Reference 32

Resolution
verified exact
local_arxiv, observed 2026-05-11T23:36:36.367470Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:a89fb945406eeafa4f357f6a72bd741b20a8b4ef7adba5a56abc457113cd85cc

Observation e3e8657e-44f7-45e8-bccc-77b9a901ca21 · outbound

This paper cites Language models are unsupervised multitask learners.OpenAI blog, 1(8):9.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Language models are unsupervised multitask learners.OpenAI blog, 1(8):9

Reference 33

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.589254Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:c5bd7ed5203c1c8e702937575378543c5de1fd02d7ae78be8e0766a42695666d

Observation 689beb14-7f54-4479-9f72-dd8082593b09 · outbound

This paper cites Pretraining task diversity and the emergence of non-bayesian in-context learning for regression.Advances in neural information processing systems, 36:14228–14246.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Pretraining task diversity and the emergence of non-bayesian in-context learning for regression.Advances in neural information processing systems, 36:14228–14246

Reference 34

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.615979Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:f072c65c72be4c7f16c0346a75fc72baebbba6f7b44a64a5ae225e19628cd40b

Observation 27aa59d6-bf86-4ec7-92df-4905a5c97e8c · outbound

This paper cites Transformers learn factored representations.arXiv preprint arXiv:2602.02385.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Transformers learn factored representations.arXiv preprint arXiv:2602.02385

Reference 35

Resolution
verified exact
arxiv_id, observed 2026-05-11T23:36:36.364038Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:8ae3f47f83d3387d79eec9e9dab87511a88d60102fc8589378decfc37541c355

Observation 3575dc73-a2b9-42d1-9a2b-39a54d8acd3c · outbound

This paper cites Transformers represent belief state geometry in their residual stream.Advances in Neural Information Processing Systems, 37:75012–75034.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Transformers represent belief state geometry in their residual stream.Advances in Neural Information Processing Systems, 37:75012–75034

Reference 36

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.571186Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:c1a30f6bf06b0e870136241a2bb9f5d63fb617166b94fca950c38ab3e1a4e0f5

Observation 2515f8af-2eef-42e0-8b3a-2c6a3cab2338 · outbound

This paper cites Uncovering hidden geometry in Transformers via disentangling position and context.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Uncovering hidden geometry in Transformers via disentangling position and context

Reference 37

Resolution
verified exact
arxiv_id, observed 2026-05-11T23:36:36.382374Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:1ea2325bb0e39b1ca7d5cd651c28dd0ef4b00b694629ad2b77a8778fec1f21c1

Observation 66a94261-7ce7-4e93-9a6c-3c1b2a17c6d2 · outbound

This paper cites Out-of-distribution generalization via composition: A lens through induction heads in transformers.Proceedings of the National Academy of Sciences, 122(6):e2417182122.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Out-of-distribution generalization via composition: A lens through induction heads in transformers.Proceedings of the National Academy of Sciences, 122(6):e2417182122

Reference 38

Resolution
verified exact
doi, observed 2026-05-09T01:54:34.543343Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:f6ce6f328790fb65ff2cdb345ba79d3ac544d6e2b29f7ea0010a5cef645bb5b9

Observation 6c04bc5c-b193-4e9d-b6b0-222cbf4ce79b · outbound

This paper cites Roformer: Enhanced transformer with rotary position embedding.Neurocomputing, 568:127063.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Roformer: Enhanced transformer with rotary position embedding.Neurocomputing, 568:127063

Reference 39

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.620783Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:c6d4499b11a605596c16a6cb40665e623876eb56b18cc0bcfb48f4433d16055a

Observation 304bba1c-806e-4c74-9c66-98960494c566 · outbound

This paper cites Daniel Freeman, Theodore R.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Daniel Freeman, Theodore R

Reference 40

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.574776Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:dda27b9d92792c083938117cdaaf3eb105a5a10672daefd4c3dab6b125f522e3

Observation 755a5894-4745-40e2-8257-782e7f9c435a · outbound

This paper cites Function vectors in large language models.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Function vectors in large language models

Reference 41

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.550130Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:31a45f0f9ce109c4b4fe11e286156caab005b4b08ae6db7c467de506b5ce74ac

Observation 7259a0f7-13d3-4f9c-96a9-5b7b699d6eb8 · outbound

This paper cites an unresolved cited work.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Unresolved cited work

Reference 42

Resolution
unresolved
raw_fallback, observed 2026-05-27T02:03:23.607825Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:5b3539fbe56cc392b2e5463ec3930fbe23780c0ca14db43402678b6a6ff9e926

Observation 6f53949b-0f0d-4ea5-8ea1-0e55153036ce · outbound

This paper cites Cambridge Series in Statistical and Probabilistic Mathematics.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Cambridge Series in Statistical and Probabilistic Mathematics

Reference 43

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.547045Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:2dde5e07fc608336fc3ea66bfd6fdc8615df84dd6081680d425566e8aededb04

Observation 305d9b09-4855-4f35-847c-490e177802b8 · outbound

This paper cites Grokking of implicit reasoning in transformers: A mechanistic journey to the edge of generalization.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Grokking of implicit reasoning in transformers: A mechanistic journey to the edge of generalization

Reference 44

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.539946Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:90a5c63814db717c97a59ace2f13491c1bb1b216f295ee217fb880ed1db0434c

Observation eccdde03-1129-4c8d-8c62-4dd1bd6165a7 · outbound

This paper cites Dif- ferentiation and specialization of attention heads via the refined local learning coefficient.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Dif- ferentiation and specialization of attention heads via the refined local learning coefficient

Reference 45

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.524473Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:7b533315c53ced417ebb9b133ac5d21183fecd232246b8e99ea5b8dc00d8c8ad

Observation 2bc6d862-5e85-4348-a4ce-8ae15f615a1c · outbound

This paper cites Larger language models do in-context learning differently.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Larger language models do in-context learning differently

Reference 46

Resolution
metadata mismatch
arxiv_id, observed 2026-05-11T23:36:36.356040Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:99d94025a077fac4b8e7d908449a49c8cc948c940dbfe29695152468d315e2a4

Observation 0042f138-9cbe-4707-b936-0cfc74be28f0 · outbound

This paper cites Transformers: State-of-the-Art Natural Language Processing.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Transformers: State-of-the-Art Natural Language Processing

Reference 47

Resolution
verified exact
doi, observed 2026-05-09T01:54:34.583497Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:ea0e3c68eabe53e6b1d0920a2663774bd9e9653741c56e6daec5b8b8dd8dfca3

Observation aac9ea5e-8d27-4942-9b22-36132b202834 · outbound

This paper cites An explanation of in-context learning as implicit bayesian inference.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers An explanation of in-context learning as implicit bayesian inference

Reference 48

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.530221Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:39fa22eace3379af559491837c527fa0454cc29475f6344f77f2e50e0edd1514

Observation c48b9019-2157-4498-ae31-a585ab2e8c66 · outbound

This paper cites number of independent directions along whichp(z)varies.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers number of independent directions along whichp(z)varies

Reference 49

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.611744Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:a5388d40575f42cd78863eb0be6ba0874024b7150cc83fb0bdae65a08b844246

Observation a7aee1e9-16f0-4557-ad84-afd9fc11c25f · outbound

This paper cites an unresolved cited work.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers Unresolved cited work

Reference 50

Resolution
unresolved
raw_fallback, observed 2026-05-27T02:03:23.534024Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:f86b393099b43cd882eccb9fe131afab9442cc7276d97c379782147af4c6c0b7

Observation 295b9d83-ee57-423c-804d-22b023277198 · outbound

This paper cites table”→“t.

Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers table”→“t

Reference 51

Resolution
verified fuzzy
raw_fallback, observed 2026-05-27T02:03:23.521408Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-03T06:30:56.289259+00:00.

source=pdf_text observed=2026-05-07T16:37:26.420806Z digest=sha256:29abeced280e1ba6dce0c98dcaea5f2d252d0639adc4943573bea5a89581fd67

Pith citing papers

No inbound Pith citation observations are available.