{"as_of":"2026-08-11T07:30:00Z","caps":{"database_statements":6,"inbound":100,"outbound":100},"context_digest":"sha256:c1e9e9867f1e66ed91b18992689e769ca9698c42968333e79fdb28b89288c8de","coverage":[{"denominator":20,"lane":"reference_resolution","note":"Typed states for the displayed outbound observations.","records_observed":20,"source":"paper_references, paper_reference_links","source_observed_at":"2026-08-10T14:49:55.605979Z","state":"measured"},{"denominator":22,"lane":"standing_notices","note":"One-hop event checks from named stored sources.","records_observed":22,"source":"scholarly_work_events, retraction_status_cache","source_observed_at":"2026-08-11T06:34:44.6726+00:00","state":"measured"},{"denominator":2,"lane":"inbound_itemization","note":"Pith citing papers itemized under the disclosed page cap.","records_observed":2,"source":"paper_references, paper_reference_links","source_observed_at":"2026-06-27T07:14:52.640891Z","state":"measured"},{"denominator":1,"lane":"external_citation_measurements","note":"A source-named dated measurement, never combined with another source.","records_observed":0,"source":"arxiv_reference","source_observed_at":"2026-07-03T14:08:21.714092Z","state":"measured"}],"external_citation_measurements":[],"inbound":[{"citation":{"cited_paper":{"arxiv_id":"2501.14925","last_updated":"2025-01-28T17:52:16Z","snapshot_observed_at":"2026-08-10T14:44:42.690129Z","submitted_at":"2025-01-24T21:31:11Z","title":"Profiling Apple Silicon Performance for ML Training","version":2},"cited_work":{"arxiv_id":"2501.14925","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2501.14925","snapshot_observed_at":"2026-07-03T14:08:21.714092Z","title":"Profiling apple silicon performance for ml training","venue":null,"work_id":"a06e10c6-a455-4eec-83c0-9c2310dbfe25","year":2025},"citing_paper":{"arxiv_id":"2502.15761","last_updated":"2026-05-12T18:27:21Z","snapshot_observed_at":"2026-08-07T01:26:25.541943Z","submitted_at":"2025-02-13T20:55:48Z","title":"AIvaluateXR: An Evaluation Framework for on-Device AI in XR with Benchmarking Results","version":3},"reference_index":59,"source":"pdf_text","source_observed_at":"2026-05-23T03:18:44.070648Z"},"links":{"cited_paper":"/paper/2501.14925","citing_paper":"/paper/2502.15761"},"observation_digest":"sha256:8a3259920e30b3263665c47fda44d671aae066baec9e0dbb758d32090815e861","observation_id":"a0c04688-99d0-4e94-82bb-f3f92669a043","resolution":{"observed_at":"2026-05-23T03:22:28.119257Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-11T06:34:44.6726+00:00","source":"crossref"},{"observed_at":"2026-08-11T06:34:36.301508+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2501.14925","last_updated":"2025-01-28T17:52:16Z","snapshot_observed_at":"2026-08-10T14:44:42.690129Z","submitted_at":"2025-01-24T21:31:11Z","title":"Profiling Apple Silicon Performance for ML Training","version":2},"cited_work":{"arxiv_id":"2501.14925","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2501.14925","snapshot_observed_at":"2026-07-03T14:08:21.714092Z","title":"Profiling apple silicon performance for ml training","venue":null,"work_id":"a06e10c6-a455-4eec-83c0-9c2310dbfe25","year":2025},"citing_paper":{"arxiv_id":"2606.12765","last_updated":"2026-06-11T00:10:23Z","snapshot_observed_at":"2026-07-06T23:51:35.904251Z","submitted_at":"2026-06-11T00:10:23Z","title":"Rigel: Reverse-Engineering the Metal 4.1 Tensor Compute Path on the Apple M4 Max GPU","version":1},"reference_index":4,"source":"pdf_text","source_observed_at":"2026-06-27T07:14:52.640891Z"},"links":{"cited_paper":"/paper/2501.14925","citing_paper":"/paper/2606.12765"},"observation_digest":"sha256:5674775a9e3fff455c3c85772cbe09cb299c9343b89b2780d4e66cbe7337d9ed","observation_id":"0657c250-0c67-4d75-b932-44fa0653cb11","resolution":{"observed_at":"2026-07-03T14:08:21.715943Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-11T06:34:44.6726+00:00","source":"crossref"},{"observed_at":"2026-08-11T06:34:36.301508+00:00","source":"retraction_watch"}],"state":"measured"}}],"links":{"evidence":"/evidence","html":"/paper/2501.14925/citation-record","integrity":"/paper/2501.14925/integrity","json":"/paper/2501.14925/citation-record.json","paper":"/paper/2501.14925"},"outbound":[{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-10T14:49:55.881792Z","title":"https://pytorch .org/get- started/previous-versions/, 2024","venue":null,"work_id":"98f24d89-a02e-44b7-b4bf-8a698fa22e29","year":2024},"citing_paper":{"arxiv_id":"2501.14925","last_updated":"2025-01-28T17:52:16Z","snapshot_observed_at":"2026-08-10T14:44:42.690129Z","submitted_at":"2025-01-24T21:31:11Z","title":"Profiling Apple Silicon Performance for ML Training","version":2},"reference_index":1,"source":"pdf_text","source_observed_at":"2026-08-10T14:49:55.522344Z"},"links":{"citing_paper":"/paper/2501.14925"},"observation_digest":"sha256:b5d7d18af669874eb1d8a3bd1d20a6fd3364dcc0a76faa62cf6c9b65ad235a03","observation_id":"20b5bd49-f4bb-4a06-b7ff-2bdc0774f6de","resolution":{"observed_at":"2026-08-10T14:49:55.886525Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-11T06:34:44.6726+00:00","source":"crossref"},{"observed_at":"2026-08-11T06:34:36.301508+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-10T14:49:55.527417Z","title":"Tyers, and Gregor Weber","venue":null,"work_id":null,"year":2020},"citing_paper":{"arxiv_id":"2501.14925","last_updated":"2025-01-28T17:52:16Z","snapshot_observed_at":"2026-08-10T14:44:42.690129Z","submitted_at":"2025-01-24T21:31:11Z","title":"Profiling Apple Silicon Performance for ML Training","version":2},"reference_index":2,"source":"pdf_text","source_observed_at":"2026-08-10T14:49:55.527417Z"},"links":{"citing_paper":"/paper/2501.14925"},"observation_digest":"sha256:da8482697ad8a94df26c3fef9420873e1540a5dc2fd7ff2d52b538a676847290","observation_id":"2f1e797e-2630-495b-bd1c-7c0a7a4102db","resolution":{"observed_at":"2026-08-10T14:49:55.527417Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-10T14:49:55.858275Z","title":"Nvidia Marketplace","venue":null,"work_id":"3d33785a-b9a6-4e5a-b283-a838e5ff9a53","year":null},"citing_paper":{"arxiv_id":"2501.14925","last_updated":"2025-01-28T17:52:16Z","snapshot_observed_at":"2026-08-10T14:44:42.690129Z","submitted_at":"2025-01-24T21:31:11Z","title":"Profiling Apple Silicon Performance for ML Training","version":2},"reference_index":3,"source":"pdf_text","source_observed_at":"2026-08-10T14:49:55.531976Z"},"links":{"citing_paper":"/paper/2501.14925"},"observation_digest":"sha256:2627d913557a7270c0e1c0c0d4785e9189ac4ae2d4757041b64b35028b837007","observation_id":"ac3d2a42-7999-4310-97c0-7263eeaac80d","resolution":{"observed_at":"2026-08-10T14:49:55.862788Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-11T06:34:44.6726+00:00","source":"crossref"},{"observed_at":"2026-08-11T06:34:36.301508+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-10T14:49:55.829910Z","title":"MLX: Efficient and flexible machine learning on apple silicon, 2023","venue":null,"work_id":"663d56ec-cf1a-4ba2-8dad-5aed68d1721e","year":2023},"citing_paper":{"arxiv_id":"2501.14925","last_updated":"2025-01-28T17:52:16Z","snapshot_observed_at":"2026-08-10T14:44:42.690129Z","submitted_at":"2025-01-24T21:31:11Z","title":"Profiling Apple Silicon Performance for ML Training","version":2},"reference_index":4,"source":"pdf_text","source_observed_at":"2026-08-10T14:49:55.541309Z"},"links":{"citing_paper":"/paper/2501.14925"},"observation_digest":"sha256:096fbfab992542d3d6fc42753429edff85cc4539cd0fb8bacfbe58139856eac5","observation_id":"d1ff587b-1f34-4bb2-b5cb-c954727956c1","resolution":{"observed_at":"2026-08-10T14:49:55.834560Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-11T06:34:44.6726+00:00","source":"crossref"},{"observed_at":"2026-08-11T06:34:36.301508+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-10T14:49:55.815825Z","title":"Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen","venue":null,"work_id":"9b7ac391-8e79-4c46-91aa-27dc326331bd","year":2021},"citing_paper":{"arxiv_id":"2501.14925","last_updated":"2025-01-28T17:52:16Z","snapshot_observed_at":"2026-08-10T14:44:42.690129Z","submitted_at":"2025-01-24T21:31:11Z","title":"Profiling Apple Silicon Performance for ML Training","version":2},"reference_index":5,"source":"pdf_text","source_observed_at":"2026-08-10T14:49:55.545882Z"},"links":{"citing_paper":"/paper/2501.14925"},"observation_digest":"sha256:dc75448991f5e465a21b6f7c106f44055ebf2d79a5d9284708d19d12ce1268df","observation_id":"a145c97b-68c2-46e3-9707-e55dda07b9fa","resolution":{"observed_at":"2026-08-10T14:49:55.820416Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-11T06:34:44.6726+00:00","source":"crossref"},{"observed_at":"2026-08-11T06:34:36.301508+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-10T14:49:55.801530Z","title":"Apple unleashes m1","venue":null,"work_id":"d0bb2139-1051-41be-8083-65ad1ca3609e","year":2020},"citing_paper":{"arxiv_id":"2501.14925","last_updated":"2025-01-28T17:52:16Z","snapshot_observed_at":"2026-08-10T14:44:42.690129Z","submitted_at":"2025-01-24T21:31:11Z","title":"Profiling Apple Silicon Performance for ML Training","version":2},"reference_index":6,"source":"pdf_text","source_observed_at":"2026-08-10T14:49:55.550423Z"},"links":{"citing_paper":"/paper/2501.14925"},"observation_digest":"sha256:b9e3481e5fdeb542e39452ec831aa9cbd61def4001812b2cad400181450f7e4a","observation_id":"690bce1f-c8a5-4c94-bfb6-318a2abbcccd","resolution":{"observed_at":"2026-08-10T14:49:55.805850Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-11T06:34:44.6726+00:00","source":"crossref"},{"observed_at":"2026-08-11T06:34:36.301508+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-10T14:49:55.786441Z","title":"Apple introduces m2 ultra","venue":null,"work_id":"be049f62-23ab-4cca-aa45-05030f7a19dc","year":2023},"citing_paper":{"arxiv_id":"2501.14925","last_updated":"2025-01-28T17:52:16Z","snapshot_observed_at":"2026-08-10T14:44:42.690129Z","submitted_at":"2025-01-24T21:31:11Z","title":"Profiling Apple Silicon Performance for ML Training","version":2},"reference_index":7,"source":"pdf_text","source_observed_at":"2026-08-10T14:49:55.555159Z"},"links":{"citing_paper":"/paper/2501.14925"},"observation_digest":"sha256:2333d80bc3b5e81b0ca05e9b2ae749b1391dca0d1adaea4134948faf29281c96","observation_id":"6a5ce8a1-d360-47c1-ac18-17d590a3446b","resolution":{"observed_at":"2026-08-10T14:49:55.791380Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-11T06:34:44.6726+00:00","source":"crossref"},{"observed_at":"2026-08-11T06:34:36.301508+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-10T14:49:55.771051Z","title":"Apple unveils m2 pro and m2 max: next-generation chips for next-level workflows","venue":null,"work_id":"4ba8d1bc-925d-4a7b-ab6b-8dcc81291d41","year":2023},"citing_paper":{"arxiv_id":"2501.14925","last_updated":"2025-01-28T17:52:16Z","snapshot_observed_at":"2026-08-10T14:44:42.690129Z","submitted_at":"2025-01-24T21:31:11Z","title":"Profiling Apple Silicon Performance for ML Training","version":2},"reference_index":8,"source":"pdf_text","source_observed_at":"2026-08-10T14:49:55.559314Z"},"links":{"citing_paper":"/paper/2501.14925"},"observation_digest":"sha256:b87a3661b6b3f401f535bd68bbde70f33fc32bc0cebda63f60d9f40b06e28768","observation_id":"3e0ea113-080b-40bf-9edf-f6f01f592af4","resolution":{"observed_at":"2026-08-10T14:49:55.775869Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-11T06:34:44.6726+00:00","source":"crossref"},{"observed_at":"2026-08-11T06:34:36.301508+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-10T14:49:55.756618Z","title":"Accelerated pytorch training on mac","venue":null,"work_id":"c6140e26-1c9b-4cb7-adf9-c67e8f7ae1e3","year":2024},"citing_paper":{"arxiv_id":"2501.14925","last_updated":"2025-01-28T17:52:16Z","snapshot_observed_at":"2026-08-10T14:44:42.690129Z","submitted_at":"2025-01-24T21:31:11Z","title":"Profiling Apple Silicon Performance for ML Training","version":2},"reference_index":9,"source":"pdf_text","source_observed_at":"2026-08-10T14:49:55.563438Z"},"links":{"citing_paper":"/paper/2501.14925"},"observation_digest":"sha256:777f69a359e67ffb9abfea5c09d14c17233007b6491a03d20d39e87ba9bc182d","observation_id":"6f202eb9-4ea4-4cb6-a81f-eafa3d57b4b8","resolution":{"observed_at":"2026-08-10T14:49:55.761499Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-11T06:34:44.6726+00:00","source":"crossref"},{"observed_at":"2026-08-11T06:34:36.301508+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-10T14:49:55.741612Z","title":"Enabling massive deep neural networks with the graphblas","venue":null,"work_id":"3a2d0929-c161-44bf-8708-5ca199136603","year":2017},"citing_paper":{"arxiv_id":"2501.14925","last_updated":"2025-01-28T17:52:16Z","snapshot_observed_at":"2026-08-10T14:44:42.690129Z","submitted_at":"2025-01-24T21:31:11Z","title":"Profiling Apple Silicon Performance for ML Training","version":2},"reference_index":10,"source":"pdf_text","source_observed_at":"2026-08-10T14:49:55.567636Z"},"links":{"citing_paper":"/paper/2501.14925"},"observation_digest":"sha256:8f0378ec3c6f20aeba73e3ddd10b8a59d77f2512bca65436d48670378d8463a7","observation_id":"8b8b6335-00d0-42a7-b8b3-784391b025de","resolution":{"observed_at":"2026-08-10T14:49:55.747019Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-11T06:34:44.6726+00:00","source":"crossref"},{"observed_at":"2026-08-11T06:34:36.301508+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-10T14:49:55.726169Z","title":null,"venue":null,"work_id":"358e63cd-d6bb-411f-bd9b-9a83a732e6c6","year":2008},"citing_paper":{"arxiv_id":"2501.14925","last_updated":"2025-01-28T17:52:16Z","snapshot_observed_at":"2026-08-10T14:44:42.690129Z","submitted_at":"2025-01-24T21:31:11Z","title":"Profiling Apple Silicon Performance for ML Training","version":2},"reference_index":11,"source":"pdf_text","source_observed_at":"2026-08-10T14:49:55.572000Z"},"links":{"citing_paper":"/paper/2501.14925"},"observation_digest":"sha256:0c7e1fc2fb3425fe142f52825bf4dfc7bac610ebe9af004d5975c1af26d53d90","observation_id":"161f7211-98d8-4bb3-bd25-ab75ffe42c31","resolution":{"observed_at":"2026-08-10T14:49:55.731131Z","resolver_source":"raw_fallback","status":"unresolved"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-11T06:34:44.6726+00:00","source":"crossref"},{"observed_at":"2026-08-11T06:34:36.301508+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-10T14:49:55.711862Z","title":"Maas, Raymond E","venue":null,"work_id":"ccd737b4-cf57-42ed-bc2f-04f59f68f5d3","year":2011},"citing_paper":{"arxiv_id":"2501.14925","last_updated":"2025-01-28T17:52:16Z","snapshot_observed_at":"2026-08-10T14:44:42.690129Z","submitted_at":"2025-01-24T21:31:11Z","title":"Profiling Apple Silicon Performance for ML Training","version":2},"reference_index":12,"source":"pdf_text","source_observed_at":"2026-08-10T14:49:55.576182Z"},"links":{"citing_paper":"/paper/2501.14925"},"observation_digest":"sha256:10c5ac8c7e0666bc298b17594d41b2a55f98368e830408a9184a9877cb229e14","observation_id":"bd5de3a6-e8a8-4c9e-98fe-bf33650a659c","resolution":{"observed_at":"2026-08-10T14:49:55.716336Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-11T06:34:44.6726+00:00","source":"crossref"},{"observed_at":"2026-08-11T06:34:36.301508+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-10T14:49:55.580403Z","title":"Marcus, Beatrice Santorini, and Mary Ann Marcinkiewicz","venue":null,"work_id":null,"year":1993},"citing_paper":{"arxiv_id":"2501.14925","last_updated":"2025-01-28T17:52:16Z","snapshot_observed_at":"2026-08-10T14:44:42.690129Z","submitted_at":"2025-01-24T21:31:11Z","title":"Profiling Apple Silicon Performance for ML Training","version":2},"reference_index":13,"source":"pdf_text","source_observed_at":"2026-08-10T14:49:55.580403Z"},"links":{"citing_paper":"/paper/2501.14925"},"observation_digest":"sha256:8e5611d66396aeb61552f145c3b818dfbd46f0644239f169ed81d0777caf2395","observation_id":"724e12cd-8c7a-48a2-88ce-c40112115b8d","resolution":{"observed_at":"2026-08-10T14:49:55.580403Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-10T14:49:55.584606Z","title":"Mixed precision training, 2018","venue":null,"work_id":null,"year":2018},"citing_paper":{"arxiv_id":"2501.14925","last_updated":"2025-01-28T17:52:16Z","snapshot_observed_at":"2026-08-10T14:44:42.690129Z","submitted_at":"2025-01-24T21:31:11Z","title":"Profiling Apple Silicon Performance for ML Training","version":2},"reference_index":14,"source":"pdf_text","source_observed_at":"2026-08-10T14:49:55.584606Z"},"links":{"citing_paper":"/paper/2501.14925"},"observation_digest":"sha256:bd8dbe37677af99cc9e4441f676ba87a1d8da66afb3bbb578988a84c7a75ab08","observation_id":"201d0557-e95d-4827-aeef-b59e1b6054e5","resolution":{"observed_at":"2026-08-10T14:49:55.584606Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2212.04356","last_updated":"2022-12-06T18:46:04Z","snapshot_observed_at":"2026-07-06T14:28:21.844826Z","submitted_at":"2022-12-06T18:46:04Z","title":"Robust Speech Recognition via Large-Scale Weak Supervision","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2212.04356","snapshot_observed_at":"2026-08-10T14:49:55.588901Z","title":"Robust speech recognition via large- scale weak supervision","venue":null,"work_id":null,"year":2022},"citing_paper":{"arxiv_id":"2501.14925","last_updated":"2025-01-28T17:52:16Z","snapshot_observed_at":"2026-08-10T14:44:42.690129Z","submitted_at":"2025-01-24T21:31:11Z","title":"Profiling Apple Silicon Performance for ML Training","version":2},"reference_index":15,"source":"pdf_text","source_observed_at":"2026-08-10T14:49:55.588901Z"},"links":{"cited_paper":"/paper/2212.04356","citing_paper":"/paper/2501.14925"},"observation_digest":"sha256:a23b2130ee0ec0f44363ddc585aded9cde9389fe67f69c8b2e4bcef30cb2f6b4","observation_id":"32a94e51-607c-42ff-8a59-9bd688dd324c","resolution":{"observed_at":"2026-08-10T14:49:55.588901Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-10T14:49:55.593482Z","title":"Language models are unsupervised multitask learners","venue":null,"work_id":null,"year":2019},"citing_paper":{"arxiv_id":"2501.14925","last_updated":"2025-01-28T17:52:16Z","snapshot_observed_at":"2026-08-10T14:44:42.690129Z","submitted_at":"2025-01-24T21:31:11Z","title":"Profiling Apple Silicon Performance for ML Training","version":2},"reference_index":16,"source":"pdf_text","source_observed_at":"2026-08-10T14:49:55.593482Z"},"links":{"citing_paper":"/paper/2501.14925"},"observation_digest":"sha256:eaa8fb9bc8d4ac0ba3793e66e34c6c4646655afa423d331891847a0202e78423","observation_id":"6c7d98ac-fefa-4660-a398-15bddb16dc31","resolution":{"observed_at":"2026-08-10T14:49:55.593482Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-10T14:49:55.597524Z","title":"{Zero-offload}: Democratizing {billion-scale} model training","venue":null,"work_id":null,"year":2021},"citing_paper":{"arxiv_id":"2501.14925","last_updated":"2025-01-28T17:52:16Z","snapshot_observed_at":"2026-08-10T14:44:42.690129Z","submitted_at":"2025-01-24T21:31:11Z","title":"Profiling Apple Silicon Performance for ML Training","version":2},"reference_index":17,"source":"pdf_text","source_observed_at":"2026-08-10T14:49:55.597524Z"},"links":{"citing_paper":"/paper/2501.14925"},"observation_digest":"sha256:b31b2140ba3aa16aaa177cdbcdf03fc68c654a6cb7e0619698e71eaf2c98e679","observation_id":"317bf093-bca7-44b4-8430-65e3454dccdb","resolution":{"observed_at":"2026-08-10T14:49:55.597524Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-10T14:49:55.601672Z","title":"Attention is all you need","venue":null,"work_id":null,"year":2017},"citing_paper":{"arxiv_id":"2501.14925","last_updated":"2025-01-28T17:52:16Z","snapshot_observed_at":"2026-08-10T14:44:42.690129Z","submitted_at":"2025-01-24T21:31:11Z","title":"Profiling Apple Silicon Performance for ML Training","version":2},"reference_index":18,"source":"pdf_text","source_observed_at":"2026-08-10T14:49:55.601672Z"},"links":{"citing_paper":"/paper/2501.14925"},"observation_digest":"sha256:d51226fc584019654fab093b93702664582c60f57b8d7e4f87f7504eceb02b7d","observation_id":"ec88e5de-404c-46e8-8d4c-22a176391ebc","resolution":{"observed_at":"2026-08-10T14:49:55.601672Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-10T14:49:55.650797Z","title":null,"venue":null,"work_id":"ad5ee14d-8d40-4190-abb4-f4391f0b4627","year":2020},"citing_paper":{"arxiv_id":"2501.14925","last_updated":"2025-01-28T17:52:16Z","snapshot_observed_at":"2026-08-10T14:44:42.690129Z","submitted_at":"2025-01-24T21:31:11Z","title":"Profiling Apple Silicon Performance for ML Training","version":2},"reference_index":19,"source":"pdf_text","source_observed_at":"2026-08-10T14:49:55.605979Z"},"links":{"citing_paper":"/paper/2501.14925"},"observation_digest":"sha256:4f5bda31ff3d16d6cf7e31235c9a67a98b98a6e5652a36cdde6f6e1fbd6c7f0b","observation_id":"9b618728-6b18-44f6-99f5-3216978d6891","resolution":{"observed_at":"2026-08-10T14:49:55.656340Z","resolver_source":"raw_fallback","status":"unresolved"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-11T06:34:44.6726+00:00","source":"crossref"},{"observed_at":"2026-08-11T06:34:36.301508+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-10T14:49:55.843978Z","title":null,"venue":null,"work_id":"c67d9cb0-d589-4339-91f0-46ad65a38ec0","year":2024},"citing_paper":{"arxiv_id":"2501.14925","last_updated":"2025-01-28T17:52:16Z","snapshot_observed_at":"2026-08-10T14:44:42.690129Z","submitted_at":"2025-01-24T21:31:11Z","title":"Profiling Apple Silicon Performance for ML Training","version":2},"reference_index":2024,"source":"pdf_text","source_observed_at":"2026-08-10T14:49:55.536781Z"},"links":{"citing_paper":"/paper/2501.14925"},"observation_digest":"sha256:9ccecd6754d3b7f2a057f07bb8b62b7af4c878d969abf37d9e8c67bbb54ead64","observation_id":"57f259d5-eae1-4ba4-9641-a682ba148061","resolution":{"observed_at":"2026-08-10T14:49:55.848821Z","resolver_source":"raw_fallback","status":"parse_uncertain"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-11T06:34:44.6726+00:00","source":"crossref"},{"observed_at":"2026-08-11T06:34:36.301508+00:00","source":"retraction_watch"}],"state":"measured"}}],"paper":{"arxiv_id":"2501.14925","last_updated":"2025-01-28T17:52:16Z","latest_version":2,"primary_category":"cs.PF","snapshot_observed_at":"2026-08-10T14:44:42.690129Z","submitted_at":"2025-01-24T21:31:11Z","title":"Profiling Apple Silicon Performance for ML Training"},"reference_resolution":{"displayed":20,"state_counts":{"malformed_identifier":0,"metadata_mismatch":0,"parse_uncertain":1,"unresolved":9,"verified_exact":0,"verified_fuzzy":10},"total_outbound_references":20},"refusal":"A citation records a reference. It does not transfer a finding from one paper to another.","schema":"pith.paper-citation-record.v1","standing_sources":[{"observed_at":"2026-08-11T06:34:44.6726+00:00","source":"crossref"},{"observed_at":"2026-08-11T06:34:36.301508+00:00","source":"retraction_watch"}],"thesis":"As of 11 August 2026, this Paper Citation Record lists 20 of 20 outbound references and 2 inbound Pith citation observations for arXiv:2501.14925."}