{"as_of":"2026-08-13T08:30:00Z","caps":{"database_statements":6,"inbound":100,"outbound":100},"context_digest":"sha256:11adfebc8d4c099b5317f88e6fd576eaec830f4c9aa725c03b204f7ace43490d","coverage":[{"denominator":39,"lane":"reference_resolution","note":"Typed states for the displayed outbound observations.","records_observed":39,"source":"paper_references, paper_reference_links","source_observed_at":"2026-08-10T15:42:21.906516Z","state":"measured"},{"denominator":39,"lane":"standing_notices","note":"One-hop event checks from named stored sources.","records_observed":39,"source":"scholarly_work_events, retraction_status_cache","source_observed_at":"2026-08-13T06:32:02.005865+00:00","state":"measured"},{"denominator":0,"lane":"inbound_itemization","note":"Pith citing papers itemized under the disclosed page cap.","records_observed":0,"source":"paper_references, paper_reference_links","source_observed_at":null,"state":"measured"},{"denominator":1,"lane":"external_citation_measurements","note":"A source-named dated measurement, never combined with another source.","records_observed":0,"source":"cited_works","source_observed_at":null,"state":"measured"}],"external_citation_measurements":[],"inbound":[],"links":{"evidence":"/evidence","html":"/paper/2608.07055/citation-record","integrity":"/paper/2608.07055/integrity","json":"/paper/2608.07055/citation-record.json","paper":"/paper/2608.07055"},"outbound":[{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-10T15:42:23.223277Z","title":null,"venue":null,"work_id":"92fa6733-074d-4ce1-b38a-947f2759c5f1","year":2024},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":1,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.668378Z"},"links":{"citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:1a581a73785c39c32a3300eef1905f623ea61f6b792494f081d049757b3b5edf","observation_id":"a6ed5389-266f-40c1-a387-e0c2174eb1dc","resolution":{"observed_at":"2026-08-10T15:42:23.228684Z","resolver_source":"raw_fallback","status":"unresolved"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-13T06:32:02.005865+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-13T06:32:02.005865+00:00","source":"crossref"},{"observed_at":"2026-08-13T06:31:53.387327+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":"2026-08-10T15:42:21.675786Z","title":null,"venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":2,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.675786Z"},"links":{"citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:95224eacf256d46758014d1815f692306bb6bc51ea8cba9757e3627853e4f68f","observation_id":"54f71b4e-4b4b-4e1b-8998-12055aa7bcfb","resolution":{"observed_at":"2026-08-10T15:42:21.675786Z","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":"2026-08-10T15:42:21.683050Z","title":null,"venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":3,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.683050Z"},"links":{"citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:155059890c4af08487bf82d5eaeb2b1656f8b7ff2c823cdc83b088900a8904a3","observation_id":"01cbf113-9469-4a4d-ae40-bd4c82e465ca","resolution":{"observed_at":"2026-08-10T15:42:21.683050Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2604.24432","last_updated":"2026-04-27T12:59:53Z","snapshot_observed_at":"2026-07-06T23:10:29.508602Z","submitted_at":"2026-04-27T12:59:53Z","title":"Kwai Summary Attention Technical Report","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2604.24432","snapshot_observed_at":"2026-08-10T15:42:21.693229Z","title":null,"venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":4,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.693229Z"},"links":{"cited_paper":"/paper/2604.24432","citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:47ecb6267bd67a3546368f810f55c801de4d96b0d67c69222f5b41ed8f057786","observation_id":"aa15c859-9e13-45da-878e-5cb610b267b0","resolution":{"observed_at":"2026-08-10T15:42:21.693229Z","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":"2026-08-10T15:42:21.699714Z","title":null,"venue":null,"work_id":null,"year":2022},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":5,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.699714Z"},"links":{"citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:b0b5a7bf73bf3fb24b77a07185406dbafc09b3384ab3377fae455c771e094dec","observation_id":"3c7279ab-427c-49bd-b02a-886efe2f5d15","resolution":{"observed_at":"2026-08-10T15:42:21.699714Z","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":"2026-08-10T15:42:23.171037Z","title":null,"venue":null,"work_id":"2ef66262-34df-43ce-9ddf-4b0d2880e32e","year":2023},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":6,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.708532Z"},"links":{"citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:40c871632207edbbd5617a08ed275c3d5d4467dbd5900f718dd34e16d23c6f19","observation_id":"f94eece9-982e-45b7-a0bd-628240676fcb","resolution":{"observed_at":"2026-08-10T15:42:23.178225Z","resolver_source":"raw_fallback","status":"unresolved"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-13T06:32:02.005865+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-13T06:32:02.005865+00:00","source":"crossref"},{"observed_at":"2026-08-13T06:31:53.387327+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2605.29755","last_updated":"2026-05-29T07:46:21Z","snapshot_observed_at":"2026-07-06T23:39:10.583797Z","submitted_at":"2026-05-28T10:59:07Z","title":"Rec-Distill: An Industrial Distillation Pipeline for Large-Scale Recommendation Models","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2605.29755","snapshot_observed_at":"2026-08-10T15:42:21.715657Z","title":null,"venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":7,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.715657Z"},"links":{"cited_paper":"/paper/2605.29755","citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:6a920f5f5ba77134d61483ce8e42449d80f29b237616e846f8f1ad05a8461fbd","observation_id":"e5ac5916-b25a-4cd5-9ed6-efc99e490ca2","resolution":{"observed_at":"2026-08-10T15:42:21.715657Z","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":"2026-08-10T15:42:21.721415Z","title":null,"venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":8,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.721415Z"},"links":{"citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:ec37bb48bd82479cff4cfea8fcc377b836e58bfa556b98779eb549d1a2c59bdb","observation_id":"b13d39c4-787a-4ee0-838a-539b3f5212e4","resolution":{"observed_at":"2026-08-10T15:42:21.721415Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2511.06077","last_updated":"2026-05-19T07:01:46Z","snapshot_observed_at":"2026-07-06T22:35:18.393401Z","submitted_at":"2025-11-08T17:22:54Z","title":"Make It Long, Keep It Fast: End-to-End 10K Long User Behavior Sequence Modeling for Billion-Scale Douyin Recommendation","version":3},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2511.06077","snapshot_observed_at":"2026-08-10T15:42:21.726663Z","title":null,"venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":9,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.726663Z"},"links":{"cited_paper":"/paper/2511.06077","citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:2054f4fe0d268cf7880fd57961bce70f7e4343c4be7a432b22629651d12227ed","observation_id":"e76f2450-0576-4f14-a790-da4c7f182d1d","resolution":{"observed_at":"2026-08-10T15:42:21.726663Z","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":"2026-08-10T15:42:21.732378Z","title":null,"venue":null,"work_id":null,"year":2022},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":10,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.732378Z"},"links":{"citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:b5b3e73381b9379baa16c1b3c511f6aa8f0545f395c0aa582632272d970fa7df","observation_id":"11947bc1-817b-48ff-99f6-2aa8c4301f8f","resolution":{"observed_at":"2026-08-10T15:42:21.732378Z","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":"2026-08-10T15:42:21.738415Z","title":null,"venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":11,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.738415Z"},"links":{"citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:66b730e14406568868dece1ab2a917ecd155c4d6604b13eeb0e7c4a3322c0edc","observation_id":"1e3ce66e-7303-4546-b81a-d4f8b354a844","resolution":{"observed_at":"2026-08-10T15:42:21.738415Z","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":"2026-08-10T15:42:21.743910Z","title":null,"venue":null,"work_id":null,"year":2021},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":12,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.743910Z"},"links":{"citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:a9bc3cecce180cd997771831a27f31f2b673a068f86a5fcfdc6881a9ea97ab38","observation_id":"0470a710-669d-4591-8b3f-13e95205c80d","resolution":{"observed_at":"2026-08-10T15:42:21.743910Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2605.29280","last_updated":"2026-06-02T18:06:36Z","snapshot_observed_at":"2026-08-04T21:09:58.253766Z","submitted_at":"2026-05-28T02:59:46Z","title":"LoopFM: Learning frOm HistOrical RePresentations of Foundation Model for Recommendation","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2605.29280","snapshot_observed_at":"2026-08-10T15:42:21.749271Z","title":null,"venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":13,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.749271Z"},"links":{"cited_paper":"/paper/2605.29280","citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:39cae29c04d8f46c5251da7d18a976ffd3c023789f8f5dbbb2e338cddecee9fb","observation_id":"18e9d28a-a172-4dcb-887e-9adf6c4c08af","resolution":{"observed_at":"2026-08-10T15:42:21.749271Z","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":"2026-08-10T15:42:23.119848Z","title":null,"venue":null,"work_id":"04bbf7f6-191b-4f9e-bd85-f59c4785e864","year":2024},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":14,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.754466Z"},"links":{"citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:541c0117be453f303f3080752fde723e0e2a16d3b844bbd41052098910cff638","observation_id":"bfd99ae3-6077-4b98-98e3-5a1a3c6cd7e5","resolution":{"observed_at":"2026-08-10T15:42:23.125391Z","resolver_source":"raw_fallback","status":"unresolved"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-13T06:32:02.005865+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-13T06:32:02.005865+00:00","source":"crossref"},{"observed_at":"2026-08-13T06:31:53.387327+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":"2026-08-10T15:42:21.759801Z","title":null,"venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":15,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.759801Z"},"links":{"citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:d5c2673e286ace25d6f0e1bd1f8f2ffc14569045744bc2dcbad4de67e6cd876d","observation_id":"45c502d9-b692-4cc1-8538-c3d42606d780","resolution":{"observed_at":"2026-08-10T15:42:21.759801Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2508.17125","last_updated":"2025-08-23T19:58:18Z","snapshot_observed_at":"2026-08-05T17:01:37.135414Z","submitted_at":"2025-08-23T19:58:18Z","title":"VQL: An End-to-End Context-Aware Vector Quantization Attention for Ultra-Long User Behavior Modeling","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2508.17125","snapshot_observed_at":"2026-08-10T15:42:21.765330Z","title":null,"venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":16,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.765330Z"},"links":{"cited_paper":"/paper/2508.17125","citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:b8182990106ffa2fc8042c434176d8395b34473defb2ac259cac313536c2adff","observation_id":"553a08a8-bbe5-46dc-94ca-b32e585ac4d4","resolution":{"observed_at":"2026-08-10T15:42:21.765330Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2604.08933","last_updated":"2026-04-10T04:02:58Z","snapshot_observed_at":"2026-08-12T21:21:00.379620Z","submitted_at":"2026-04-10T04:02:58Z","title":"IAT: Instance-As-Token Compression for Historical User Sequence Modeling in Industrial Recommender Systems","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2604.08933","snapshot_observed_at":"2026-08-10T15:42:21.772049Z","title":null,"venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":17,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.772049Z"},"links":{"cited_paper":"/paper/2604.08933","citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:344791327fa195bd2be8dce686198206d7a174b81c9244eb5a305f9ed9d9234a","observation_id":"72c9b625-bc1f-4059-81e5-d0fb6c671fe6","resolution":{"observed_at":"2026-08-10T15:42:21.772049Z","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":"2026-08-10T15:42:21.779460Z","title":null,"venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":18,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.779460Z"},"links":{"citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:d4bda0a464a799a0b10f7d147c083dd9a6ba2a6cb244042880a684670adad1f7","observation_id":"d7370439-8dba-49d8-99a8-fa3402a431c6","resolution":{"observed_at":"2026-08-10T15:42:21.779460Z","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":"2026-08-10T15:42:21.785597Z","title":null,"venue":null,"work_id":null,"year":2021},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":19,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.785597Z"},"links":{"citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:d429d85cf89da2374d7f27c348b52c68ba287159119442c433c9b9d54f1a17d2","observation_id":"93280cd2-ccf6-4220-a970-c23d04d061af","resolution":{"observed_at":"2026-08-10T15:42:21.785597Z","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":"2026-08-10T15:42:21.791598Z","title":null,"venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":20,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.791598Z"},"links":{"citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:638ae78e17ad1033e22459fc3bb65a7164559736bdfa3e32aada60065f90e2c3","observation_id":"4fedbe98-2cc2-429e-80fd-3c45a0ccd207","resolution":{"observed_at":"2026-08-10T15:42:21.791598Z","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":"2026-08-10T15:42:21.800250Z","title":null,"venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":21,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.800250Z"},"links":{"citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:d13b0a5159e0e8debc2ae520c3bd4981bced21af778722c539a06b6175d5bdb6","observation_id":"59cd97be-0f1a-485b-ab12-9c2076b27570","resolution":{"observed_at":"2026-08-10T15:42:21.800250Z","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":"2026-08-10T15:42:21.806089Z","title":null,"venue":null,"work_id":null,"year":2020},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":22,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.806089Z"},"links":{"citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:fd744bbab31f2c254a91c08c8355d724c1290b3890ab4f1d6768b8bf57f059ef","observation_id":"b4ea440b-113a-45e8-9683-493fcf336a3c","resolution":{"observed_at":"2026-08-10T15:42:21.806089Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2502.17494","last_updated":"2025-07-14T03:01:53Z","snapshot_observed_at":"2026-08-07T18:00:31.644042Z","submitted_at":"2025-02-20T22:35:52Z","title":"External Large Foundation Model: How to Efficiently Serve Trillions of Parameters for Online Ads Recommendation","version":7},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2502.17494","snapshot_observed_at":"2026-08-10T15:42:21.811030Z","title":null,"venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":23,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.811030Z"},"links":{"cited_paper":"/paper/2502.17494","citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:52734923900fabb61cd2c26a7ea5c7cb1b79f091ba34f79afb16232bd588aa7e","observation_id":"b4577454-96a9-4728-88c8-5937317652eb","resolution":{"observed_at":"2026-08-10T15:42:21.811030Z","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":"2026-08-10T15:42:21.816453Z","title":null,"venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":24,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.816453Z"},"links":{"citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:e9dc5cb8635f074c3701648efa6f78218d701a7344f87853a0d3b4ff28eba910","observation_id":"2e8fbb06-30af-48aa-9436-9eabedc582b9","resolution":{"observed_at":"2026-08-10T15:42:21.816453Z","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":"2602.11235","doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-10T15:42:22.418502Z","title":null,"venue":null,"work_id":"d7c0b818-5158-48ba-a095-d36118d6bb65","year":2026},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":25,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.821745Z"},"links":{"citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:8cb2dfa4b0b1c9962e20bb953965c58274f8b7c6ae249bb2b16b5f4c67c2e549","observation_id":"490f5d9d-86b4-41d3-9816-b1c7f3bf0155","resolution":{"observed_at":"2026-08-10T15:42:22.428367Z","resolver_source":"raw_fallback","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-13T06:32:02.005865+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-13T06:32:02.005865+00:00","source":"crossref"},{"observed_at":"2026-08-13T06:31:53.387327+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":"2026-08-10T15:42:21.827448Z","title":null,"venue":null,"work_id":null,"year":2017},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":26,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.827448Z"},"links":{"citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:b75ddc8ae04796e526cb7dee4a5cecbd7a23873af3ed5f87c9522dbfa67861cc","observation_id":"d4720c32-a1e3-44bd-a448-e044adddbbb1","resolution":{"observed_at":"2026-08-10T15:42:21.827448Z","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":"2026-08-10T15:42:23.016821Z","title":null,"venue":null,"work_id":"0ea934e6-4059-466a-9cb3-61e0576baae1","year":2019},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":27,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.833042Z"},"links":{"citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:ebb874d315d1eec884998aa0e0de599cf0f8439612fd5d0fa9154b0bac4eb5b3","observation_id":"d470733a-5528-4ce5-89da-df2e27e17c5a","resolution":{"observed_at":"2026-08-10T15:42:23.022150Z","resolver_source":"raw_fallback","status":"unresolved"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-13T06:32:02.005865+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-13T06:32:02.005865+00:00","source":"crossref"},{"observed_at":"2026-08-13T06:31:53.387327+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2604.15650","last_updated":"2026-07-16T11:34:49Z","snapshot_observed_at":"2026-08-11T15:25:00.459239Z","submitted_at":"2026-04-17T02:47:31Z","title":"Sample Is Feature: Beyond Item-Level, Toward Sample-Level Tokens for Unified Large Recommender Models","version":3},"cited_work":{"arxiv_id":"2604.15650","doi":null,"metadata_source":"pith","pith_arxiv_id":"2604.15650","snapshot_observed_at":"2026-08-10T15:42:22.278162Z","title":"Sample Is Feature: Beyond Item-Level, Toward Sample-Level Tokens for Unified Large Recommender Models","venue":"cs.IR","work_id":"02330871-158d-44d5-aa19-f093a1ed84ba","year":2026},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":28,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.838585Z"},"links":{"cited_paper":"/paper/2604.15650","citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:310f5b5cf5925dc7ae69e809f674611bba0485a06c86850c88ded41671dd0777","observation_id":"12e2688d-a92a-4dc2-a71c-88d057e988c6","resolution":{"observed_at":"2026-08-10T15:42:22.285608Z","resolver_source":"local_arxiv","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-13T06:32:02.005865+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-13T06:32:02.005865+00:00","source":"crossref"},{"observed_at":"2026-08-13T06:31:53.387327+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":"2026-08-10T15:42:21.844059Z","title":null,"venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":29,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.844059Z"},"links":{"citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:c3022b6b229e97ca17c169fe2087a6c5c4b216f32d673b2e1ee011863f2cf26d","observation_id":"8ad03a8a-1e9a-4921-995d-9e2ecc75ddaf","resolution":{"observed_at":"2026-08-10T15:42:21.844059Z","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":"2026-08-10T15:42:21.850155Z","title":null,"venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":30,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.850155Z"},"links":{"citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:7d23b201c8ff4148d07229b2d936478eb3a084299f7333da1d1ce90f5aae2308","observation_id":"3e284b5d-0347-4da3-ba8e-2c514d5ec734","resolution":{"observed_at":"2026-08-10T15:42:21.850155Z","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":"2026-08-10T15:42:21.855403Z","title":null,"venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":31,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.855403Z"},"links":{"citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:bdf83ce2185753a89d6ed11f5a9780aa307cd71bc93199dee2dfeb0358211181","observation_id":"e91065b4-4367-4d6a-85ce-86294ffd45a4","resolution":{"observed_at":"2026-08-10T15:42:21.855403Z","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":"2026-08-10T15:42:21.860496Z","title":null,"venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":32,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.860496Z"},"links":{"citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:0901739dbcfa82eb8593922eade38d0417746640a226729c0d67c55d335cffce","observation_id":"88641728-22a2-4e7a-95ba-200fc580a391","resolution":{"observed_at":"2026-08-10T15:42:21.860496Z","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":"2026-08-10T15:42:21.870143Z","title":null,"venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":33,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.870143Z"},"links":{"citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:45cca8ee6109bde78650ca0bce77c3d3f0177e83534722d519e66a08af78b82a","observation_id":"86ca2628-e9c3-4df7-b5a1-97178981b9e3","resolution":{"observed_at":"2026-08-10T15:42:21.870143Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2602.21009","last_updated":"2026-07-03T11:52:26Z","snapshot_observed_at":"2026-08-10T03:48:55.360340Z","submitted_at":"2026-02-24T15:28:58Z","title":"HiSAC: Hierarchical Sparse Activation Compression for Ultra-long Sequence Modeling in Recommenders","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2602.21009","snapshot_observed_at":"2026-08-10T15:42:21.875737Z","title":null,"venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":34,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.875737Z"},"links":{"cited_paper":"/paper/2602.21009","citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:2fc1b203927ba5452fbfca0b1fddb8e8e28c1958fb97f2bfab21ae64ef2587bb","observation_id":"d9d8f296-75fa-4349-a0b5-e30ecd495439","resolution":{"observed_at":"2026-08-10T15:42:21.875737Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2402.17152","last_updated":"2024-05-06T02:05:45Z","snapshot_observed_at":"2026-08-10T13:07:04.859020Z","submitted_at":"2024-02-27T02:37:37Z","title":"Actions Speak Louder than Words: Trillion-Parameter Sequential Transducers for Generative Recommendations","version":3},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2402.17152","snapshot_observed_at":"2026-08-10T15:42:21.881862Z","title":null,"venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":35,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.881862Z"},"links":{"cited_paper":"/paper/2402.17152","citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:85b38ac2f177d7833079a82ce79bf96bb66d4de9cb786788711af9c5095802b2","observation_id":"b5a0eab4-6509-4d05-98de-26e8332a7d78","resolution":{"observed_at":"2026-08-10T15:42:21.881862Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2508.14948","last_updated":"2025-08-20T10:18:01Z","snapshot_observed_at":"2026-08-09T13:49:56.588898Z","submitted_at":"2025-08-20T10:18:01Z","title":"Large Foundation Model for Ads Recommendation","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2508.14948","snapshot_observed_at":"2026-08-10T15:42:21.887390Z","title":null,"venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":36,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.887390Z"},"links":{"cited_paper":"/paper/2508.14948","citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:1cea4af0331272aa5b9f0deaae60aacedf2b1116cd28efdda6bc53728ca77f61","observation_id":"67641885-8cf0-480b-822e-7618da32e8f0","resolution":{"observed_at":"2026-08-10T15:42:21.887390Z","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":"2026-08-10T15:42:21.894403Z","title":null,"venue":null,"work_id":null,"year":2019},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":37,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.894403Z"},"links":{"citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:321f11dfa990d6cbd606b15383697053586dd72478c6dc37b9b26370f870c12b","observation_id":"d4b26531-b85e-458b-a163-28388b5e5833","resolution":{"observed_at":"2026-08-10T15:42:21.894403Z","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":"2026-08-10T15:42:21.899527Z","title":null,"venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":38,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.899527Z"},"links":{"citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:fb11b75be250776708e8981cd4bccf8ba2f55737771d2a87f106266a41d33748","observation_id":"fb998209-ccda-4eaf-b3b3-fa5c31a2656b","resolution":{"observed_at":"2026-08-10T15:42:21.899527Z","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":"2026-08-10T15:42:22.825498Z","title":null,"venue":null,"work_id":"89b31f72-eca4-48ec-9ae7-3f38c2570fe2","year":2025},"citing_paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation","version":1},"reference_index":39,"source":"pdf_text","source_observed_at":"2026-08-10T15:42:21.906516Z"},"links":{"citing_paper":"/paper/2608.07055"},"observation_digest":"sha256:c289b72230a0c31b83afde2ee53850a906c5d093d0b8d66414c30805c6a4a2e4","observation_id":"5d8ecb26-3cae-46e2-aa2c-08f615407f1f","resolution":{"observed_at":"2026-08-10T15:42:22.831082Z","resolver_source":"raw_fallback","status":"unresolved"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-13T06:32:02.005865+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-13T06:32:02.005865+00:00","source":"crossref"},{"observed_at":"2026-08-13T06:31:53.387327+00:00","source":"retraction_watch"}],"state":"measured"}}],"paper":{"arxiv_id":"2608.07055","last_updated":"2026-08-07T10:04:44Z","latest_version":1,"primary_category":"cs.IR","snapshot_observed_at":"2026-08-12T23:11:47.260552Z","submitted_at":"2026-08-07T10:04:44Z","title":"Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation"},"reference_resolution":{"displayed":39,"state_counts":{"malformed_identifier":0,"metadata_mismatch":0,"parse_uncertain":0,"unresolved":37,"verified_exact":2,"verified_fuzzy":0},"total_outbound_references":39},"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":"2026-08-13T06:32:02.005865+00:00","source":"crossref"},{"observed_at":"2026-08-13T06:31:53.387327+00:00","source":"retraction_watch"}],"thesis":"As of 13 August 2026, this Paper Citation Record lists 39 of 39 outbound references and 0 inbound Pith citation observations for arXiv:2608.07055."}