{"as_of":"2026-08-10T18:07:00Z","caps":{"database_statements":6,"inbound":100,"outbound":100},"context_digest":"sha256:258b3f901a56992d67a52565a8b6a1df610e1d02b536ada2f5c4937870d10911","coverage":[{"denominator":0,"lane":"reference_resolution","note":"Typed states for the displayed outbound observations.","records_observed":0,"source":"paper_references, paper_reference_links","source_observed_at":null,"state":"measured"},{"denominator":23,"lane":"standing_notices","note":"One-hop event checks from named stored sources.","records_observed":23,"source":"scholarly_work_events, retraction_status_cache","source_observed_at":"2026-08-10T06:31:04.303077+00:00","state":"measured"},{"denominator":23,"lane":"inbound_itemization","note":"Pith citing papers itemized under the disclosed page cap.","records_observed":23,"source":"paper_references, paper_reference_links","source_observed_at":"2026-08-09T17:57:25.837531Z","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":"2026-07-04T13:19:50.449132Z","state":"measured"}],"external_citation_measurements":[],"inbound":[{"citation":{"cited_paper":{"arxiv_id":"2004.12993","last_updated":"2020-04-27T17:58:05Z","snapshot_observed_at":"2026-08-06T01:22:26.725556Z","submitted_at":"2020-04-27T17:58:05Z","title":"DeeBERT: Dynamic Early Exiting for Accelerating BERT Inference","version":1},"cited_work":{"arxiv_id":"2004.12993","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2004.12993","snapshot_observed_at":"2026-07-04T13:19:50.449132Z","title":"DeeBERT: Dynam ic Early Exiting for Accelerating BERT Inference","venue":null,"work_id":"1d9ce0aa-b85e-4fca-8bde-cc21a6e6e0f2","year":2020},"citing_paper":{"arxiv_id":"2303.08112","last_updated":"2025-11-11T01:13:28Z","snapshot_observed_at":"2026-07-06T02:11:23.670680Z","submitted_at":"2023-03-14T17:47:09Z","title":"Eliciting Latent Predictions from Transformers with the Tuned Lens","version":6},"reference_index":90,"source":"arxiv_source","source_observed_at":"2026-05-12T16:54:37.382049Z"},"links":{"cited_paper":"/paper/2004.12993","citing_paper":"/paper/2303.08112"},"observation_digest":"sha256:7975276ad633a84344c164522325094c0d11d7ab7d952723509ebf902d708dfe","observation_id":"f616e550-5364-42d8-80c6-7dfc5258ceb2","resolution":{"observed_at":"2026-05-12T16:54:37.545480Z","resolver_source":"arxiv_id","status":"metadata_mismatch"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-10T06:31:04.303077+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-10T06:31:04.303077+00:00","source":"crossref"},{"observed_at":"2026-08-10T06:30:57.382061+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2004.12993","last_updated":"2020-04-27T17:58:05Z","snapshot_observed_at":"2026-08-06T01:22:26.725556Z","submitted_at":"2020-04-27T17:58:05Z","title":"DeeBERT: Dynamic Early Exiting for Accelerating BERT Inference","version":1},"cited_work":{"arxiv_id":"2004.12993","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2004.12993","snapshot_observed_at":"2026-07-04T13:19:50.449132Z","title":"DeeBERT: Dynam ic Early Exiting for Accelerating BERT Inference","venue":null,"work_id":"1d9ce0aa-b85e-4fca-8bde-cc21a6e6e0f2","year":2020},"citing_paper":{"arxiv_id":"2410.04941","last_updated":"2026-05-18T08:56:27Z","snapshot_observed_at":"2026-08-05T10:18:09.103385Z","submitted_at":"2024-10-07T11:35:24Z","title":"TOAST: Transformer Optimization using Adaptive and Simple Transformations","version":7},"reference_index":41,"source":"arxiv_source","source_observed_at":"2026-05-23T19:46:07.124996Z"},"links":{"cited_paper":"/paper/2004.12993","citing_paper":"/paper/2410.04941"},"observation_digest":"sha256:e83d72014ecfb1ef58d892f52cd8ca840f8072975108c7e01372b02e75a48886","observation_id":"703f5c42-52c4-4c34-92b3-8df30b45ec39","resolution":{"observed_at":"2026-05-23T19:48:23.054063Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-10T06:31:04.303077+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-10T06:31:04.303077+00:00","source":"crossref"},{"observed_at":"2026-08-10T06:30:57.382061+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2004.12993","last_updated":"2020-04-27T17:58:05Z","snapshot_observed_at":"2026-08-06T01:22:26.725556Z","submitted_at":"2020-04-27T17:58:05Z","title":"DeeBERT: Dynamic Early Exiting for Accelerating BERT Inference","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2004.12993","snapshot_observed_at":"2026-08-09T17:57:25.837531Z","title":null,"venue":null,"work_id":null,"year":2019},"citing_paper":{"arxiv_id":"2502.00745","last_updated":"2025-02-02T10:35:19Z","snapshot_observed_at":"2026-08-10T07:30:31.821573Z","submitted_at":"2025-02-02T10:35:19Z","title":"BEEM: Boosting Performance of Early Exit DNNs using Multi-Exit Classifiers as Experts","version":1},"reference_index":28,"source":"pdf_text","source_observed_at":"2026-08-09T17:57:25.837531Z"},"links":{"cited_paper":"/paper/2004.12993","citing_paper":"/paper/2502.00745"},"observation_digest":"sha256:e652fe77d03205240b2f6d649106ba11f6f64bb77e8906ba9ede6be2ab1234b1","observation_id":"10c0ec90-5f5e-44bc-a123-54a6a7c29592","resolution":{"observed_at":"2026-08-09T17:57:25.837531Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2004.12993","last_updated":"2020-04-27T17:58:05Z","snapshot_observed_at":"2026-08-06T01:22:26.725556Z","submitted_at":"2020-04-27T17:58:05Z","title":"DeeBERT: Dynamic Early Exiting for Accelerating BERT Inference","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2004.12993","snapshot_observed_at":"2026-08-07T05:53:46.089881Z","title":null,"venue":null,"work_id":null,"year":2020},"citing_paper":{"arxiv_id":"2506.06884","last_updated":"2025-06-07T18:26:58Z","snapshot_observed_at":"2026-08-08T01:16:34.931310Z","submitted_at":"2025-06-07T18:26:58Z","title":"FREE: Fast and Robust Vision Language Models with Early Exits","version":1},"reference_index":64,"source":"arxiv_source","source_observed_at":"2026-08-07T05:53:46.089881Z"},"links":{"cited_paper":"/paper/2004.12993","citing_paper":"/paper/2506.06884"},"observation_digest":"sha256:2774efc822da21c9e6c61763d139c824ef30845e4900dfcc5048b63b9389a0d7","observation_id":"a40d33dd-baa5-44a5-86fc-7dd146285b79","resolution":{"observed_at":"2026-08-07T05:53:46.089881Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2004.12993","last_updated":"2020-04-27T17:58:05Z","snapshot_observed_at":"2026-08-06T01:22:26.725556Z","submitted_at":"2020-04-27T17:58:05Z","title":"DeeBERT: Dynamic Early Exiting for Accelerating BERT Inference","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2004.12993","snapshot_observed_at":"2026-08-06T21:11:18.509710Z","title":null,"venue":null,"work_id":null,"year":2020},"citing_paper":{"arxiv_id":"2507.01079","last_updated":"2025-07-01T15:12:14Z","snapshot_observed_at":"2026-08-10T01:01:27.245436Z","submitted_at":"2025-07-01T15:12:14Z","title":"MobileRAG: A Fast, Memory-Efficient, and Energy-Efficient Method for On-Device RAG","version":1},"reference_index":57,"source":"pdf_text","source_observed_at":"2026-08-06T21:11:18.509710Z"},"links":{"cited_paper":"/paper/2004.12993","citing_paper":"/paper/2507.01079"},"observation_digest":"sha256:9f4b4efc090d118b06383fd7f358ea11a15a8c18f549173c2807c63665d65135","observation_id":"edd8cf25-2ddd-4d96-a5ce-13456f966539","resolution":{"observed_at":"2026-08-06T21:11:18.509710Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2004.12993","last_updated":"2020-04-27T17:58:05Z","snapshot_observed_at":"2026-08-06T01:22:26.725556Z","submitted_at":"2020-04-27T17:58:05Z","title":"DeeBERT: Dynamic Early Exiting for Accelerating BERT Inference","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2004.12993","snapshot_observed_at":"2026-08-06T18:33:20.378873Z","title":"Deebert: Dynamic early exiting for accelerating bert inference","venue":null,"work_id":null,"year":2004},"citing_paper":{"arxiv_id":"2507.07996","last_updated":"2025-07-10T17:59:53Z","snapshot_observed_at":"2026-08-10T07:07:45.618793Z","submitted_at":"2025-07-10T17:59:53Z","title":"Skip a Layer or Loop it? Test-Time Depth Adaptation of Pretrained LLMs","version":1},"reference_index":23,"source":"pdf_text","source_observed_at":"2026-08-06T18:33:20.378873Z"},"links":{"cited_paper":"/paper/2004.12993","citing_paper":"/paper/2507.07996"},"observation_digest":"sha256:5b2f8d1014268b7dc8fab6843023acc5aa0e5897e850c140b25061bd8e9528af","observation_id":"1ad0b72b-8ed4-48f1-8c13-e706fae5e322","resolution":{"observed_at":"2026-08-06T18:33:20.378873Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2004.12993","last_updated":"2020-04-27T17:58:05Z","snapshot_observed_at":"2026-08-06T01:22:26.725556Z","submitted_at":"2020-04-27T17:58:05Z","title":"DeeBERT: Dynamic Early Exiting for Accelerating BERT Inference","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2004.12993","snapshot_observed_at":"2026-08-05T21:16:26.875019Z","title":"Deebert: Dynamic early exiting for accelerating bert inference","venue":null,"work_id":null,"year":2004},"citing_paper":{"arxiv_id":"2508.09262","last_updated":"2025-08-12T18:05:33Z","snapshot_observed_at":"2026-08-08T09:26:18.699272Z","submitted_at":"2025-08-12T18:05:33Z","title":"Harnessing Input-Adaptive Inference for Efficient VLN","version":1},"reference_index":61,"source":"pdf_text","source_observed_at":"2026-08-05T21:16:26.875019Z"},"links":{"cited_paper":"/paper/2004.12993","citing_paper":"/paper/2508.09262"},"observation_digest":"sha256:06afc62ff65f076789e1976cf8e16c444d63fe021585d0c1b773cbf342b9da16","observation_id":"ae3fcb53-9d36-4050-ba7a-e5247cbe5855","resolution":{"observed_at":"2026-08-05T21:16:26.875019Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2004.12993","last_updated":"2020-04-27T17:58:05Z","snapshot_observed_at":"2026-08-06T01:22:26.725556Z","submitted_at":"2020-04-27T17:58:05Z","title":"DeeBERT: Dynamic Early Exiting for Accelerating BERT Inference","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2004.12993","snapshot_observed_at":"2026-08-04T19:27:20.989278Z","title":"Deebert: Dynamic early exiting for accelerating bert inference.arXiv preprint arXiv:2004.12993, 2020","venue":null,"work_id":null,"year":2004},"citing_paper":{"arxiv_id":"2509.09286","last_updated":"2025-09-11T09:22:16Z","snapshot_observed_at":"2026-08-08T05:44:51.958215Z","submitted_at":"2025-09-11T09:22:16Z","title":"Visual Programmability: A Guide for Code-as-Thought in Chart Understanding","version":1},"reference_index":56,"source":"pdf_text","source_observed_at":"2026-08-04T19:27:20.989278Z"},"links":{"cited_paper":"/paper/2004.12993","citing_paper":"/paper/2509.09286"},"observation_digest":"sha256:9790827a6d2e81cc3ccf92aad38265e20fb891a2e0532bea4a15bce826b95967","observation_id":"b6a73c93-bdb9-477d-9aa7-ddfe21a46b85","resolution":{"observed_at":"2026-08-04T19:27:20.989278Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2004.12993","last_updated":"2020-04-27T17:58:05Z","snapshot_observed_at":"2026-08-06T01:22:26.725556Z","submitted_at":"2020-04-27T17:58:05Z","title":"DeeBERT: Dynamic Early Exiting for Accelerating BERT Inference","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2004.12993","snapshot_observed_at":"2026-08-04T14:57:00.268518Z","title":"Deebert: Dynamic early exiting for accelerating bert inference","venue":null,"work_id":null,"year":2004},"citing_paper":{"arxiv_id":"2509.22992","last_updated":"2026-08-03T08:38:00Z","snapshot_observed_at":"2026-08-10T00:59:10.624197Z","submitted_at":"2025-09-26T23:08:03Z","title":"T-TAMER: Provably Taming Trade-offs in ML Serving","version":2},"reference_index":96,"source":"arxiv_source","source_observed_at":"2026-08-04T14:57:00.268518Z"},"links":{"cited_paper":"/paper/2004.12993","citing_paper":"/paper/2509.22992"},"observation_digest":"sha256:08b73cf45f824815bd899b53809975461eccca91d8d63c0e36b5cf44624b3422","observation_id":"d299826c-9f13-4a08-b427-dffafcbfe339","resolution":{"observed_at":"2026-08-04T14:57:00.268518Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2004.12993","last_updated":"2020-04-27T17:58:05Z","snapshot_observed_at":"2026-08-06T01:22:26.725556Z","submitted_at":"2020-04-27T17:58:05Z","title":"DeeBERT: Dynamic Early Exiting for Accelerating BERT Inference","version":1},"cited_work":{"arxiv_id":"2004.12993","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2004.12993","snapshot_observed_at":"2026-07-04T13:19:50.449132Z","title":"DeeBERT: Dynam ic Early Exiting for Accelerating BERT Inference","venue":null,"work_id":"1d9ce0aa-b85e-4fca-8bde-cc21a6e6e0f2","year":2020},"citing_paper":{"arxiv_id":"2510.12773","last_updated":"2026-05-18T21:02:06Z","snapshot_observed_at":"2026-08-03T04:39:11.410726Z","submitted_at":"2025-10-14T17:51:26Z","title":"Dr.LLM: Dynamic Layer Routing in LLMs","version":2},"reference_index":19,"source":"pdf_text","source_observed_at":"2026-05-21T20:01:48.806709Z"},"links":{"cited_paper":"/paper/2004.12993","citing_paper":"/paper/2510.12773"},"observation_digest":"sha256:78ccb39bc02344dcb600a4ac7da84bbde7d843931408f33cd771803430208733","observation_id":"e2514974-4b09-4721-8f51-c78259ff6103","resolution":{"observed_at":"2026-05-21T20:04:20.339326Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-10T06:31:04.303077+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-10T06:31:04.303077+00:00","source":"crossref"},{"observed_at":"2026-08-10T06:30:57.382061+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2004.12993","last_updated":"2020-04-27T17:58:05Z","snapshot_observed_at":"2026-08-06T01:22:26.725556Z","submitted_at":"2020-04-27T17:58:05Z","title":"DeeBERT: Dynamic Early Exiting for Accelerating BERT Inference","version":1},"cited_work":{"arxiv_id":"2004.12993","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2004.12993","snapshot_observed_at":"2026-07-04T13:19:50.449132Z","title":"DeeBERT: Dynam ic Early Exiting for Accelerating BERT Inference","venue":null,"work_id":"1d9ce0aa-b85e-4fca-8bde-cc21a6e6e0f2","year":2020},"citing_paper":{"arxiv_id":"2604.18592","last_updated":"2026-03-27T15:27:58Z","snapshot_observed_at":"2026-08-05T01:52:00.445011Z","submitted_at":"2026-03-27T15:27:58Z","title":"Two-dimensional early exit optimisation of LLM inference","version":1},"reference_index":25,"source":"pdf_text","source_observed_at":"2026-05-14T22:59:07.065193Z"},"links":{"cited_paper":"/paper/2004.12993","citing_paper":"/paper/2604.18592"},"observation_digest":"sha256:91ccd94f7fe3f9081629895338193db5f6bf4a738348fb80165f4b8a10ae739f","observation_id":"0e4a7501-f32b-470c-a47d-18f3c58e7e9b","resolution":{"observed_at":"2026-05-14T22:59:34.202511Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-10T06:31:04.303077+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-10T06:31:04.303077+00:00","source":"crossref"},{"observed_at":"2026-08-10T06:30:57.382061+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2004.12993","last_updated":"2020-04-27T17:58:05Z","snapshot_observed_at":"2026-08-06T01:22:26.725556Z","submitted_at":"2020-04-27T17:58:05Z","title":"DeeBERT: Dynamic Early Exiting for Accelerating BERT Inference","version":1},"cited_work":{"arxiv_id":"2004.12993","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2004.12993","snapshot_observed_at":"2026-07-04T13:19:50.449132Z","title":"DeeBERT: Dynam ic Early Exiting for Accelerating BERT Inference","venue":null,"work_id":"1d9ce0aa-b85e-4fca-8bde-cc21a6e6e0f2","year":2020},"citing_paper":{"arxiv_id":"2604.22906","last_updated":"2026-04-24T16:56:53Z","snapshot_observed_at":"2026-08-07T20:25:38.013655Z","submitted_at":"2026-04-24T16:56:53Z","title":"Network Edge Inference for Large Language Models: Principles, Techniques, and Opportunities","version":1},"reference_index":174,"source":"pdf_text","source_observed_at":"2026-05-08T09:45:57.201837Z"},"links":{"cited_paper":"/paper/2004.12993","citing_paper":"/paper/2604.22906"},"observation_digest":"sha256:931d403b45acc22b5789cdbaed23bfef72a5ee3c8d5f40ea523160bb94120d8c","observation_id":"1e3502b5-1fba-46f4-9d73-e01985ed0c8f","resolution":{"observed_at":"2026-05-11T20:16:10.368141Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-10T06:31:04.303077+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-10T06:31:04.303077+00:00","source":"crossref"},{"observed_at":"2026-08-10T06:30:57.382061+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2004.12993","last_updated":"2020-04-27T17:58:05Z","snapshot_observed_at":"2026-08-06T01:22:26.725556Z","submitted_at":"2020-04-27T17:58:05Z","title":"DeeBERT: Dynamic Early Exiting for Accelerating BERT Inference","version":1},"cited_work":{"arxiv_id":"2004.12993","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2004.12993","snapshot_observed_at":"2026-07-04T13:19:50.449132Z","title":"DeeBERT: Dynam ic Early Exiting for Accelerating BERT Inference","venue":null,"work_id":"1d9ce0aa-b85e-4fca-8bde-cc21a6e6e0f2","year":2020},"citing_paper":{"arxiv_id":"2604.24810","last_updated":"2026-05-22T08:16:07Z","snapshot_observed_at":"2026-07-06T23:10:47.277692Z","submitted_at":"2026-04-27T08:51:44Z","title":"A Comparative Analysis on the Performance of Upper Confidence Bound Algorithms in Adaptive Deep Neural Networks","version":2},"reference_index":19,"source":"pdf_text","source_observed_at":"2026-05-08T04:25:47.495730Z"},"links":{"cited_paper":"/paper/2004.12993","citing_paper":"/paper/2604.24810"},"observation_digest":"sha256:aa19cb490d13b029074c987fe3197e03894abd1122ebcd69894477386d7bbcf3","observation_id":"2bdab8cb-bdb3-471d-8d85-2ad6fca39186","resolution":{"observed_at":"2026-05-11T21:46:40.913935Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-10T06:31:04.303077+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-10T06:31:04.303077+00:00","source":"crossref"},{"observed_at":"2026-08-10T06:30:57.382061+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2004.12993","last_updated":"2020-04-27T17:58:05Z","snapshot_observed_at":"2026-08-06T01:22:26.725556Z","submitted_at":"2020-04-27T17:58:05Z","title":"DeeBERT: Dynamic Early Exiting for Accelerating BERT Inference","version":1},"cited_work":{"arxiv_id":"2004.12993","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2004.12993","snapshot_observed_at":"2026-07-04T13:19:50.449132Z","title":"DeeBERT: Dynam ic Early Exiting for Accelerating BERT Inference","venue":null,"work_id":"1d9ce0aa-b85e-4fca-8bde-cc21a6e6e0f2","year":2020},"citing_paper":{"arxiv_id":"2604.24810","last_updated":"2026-05-22T08:16:07Z","snapshot_observed_at":"2026-07-06T23:10:47.277692Z","submitted_at":"2026-04-27T08:51:44Z","title":"A Comparative Analysis on the Performance of Upper Confidence Bound Algorithms in Adaptive Deep Neural Networks","version":3},"reference_index":19,"source":"pdf_text","source_observed_at":"2026-05-25T06:49:44.564535Z"},"links":{"cited_paper":"/paper/2004.12993","citing_paper":"/paper/2604.24810"},"observation_digest":"sha256:6a79a2f05f12be31ff8ddb673091aad54f1f84da55726ebb93330e0e84a671a2","observation_id":"ff6bf392-2166-429b-80d4-b6a997abacfd","resolution":{"observed_at":"2026-05-25T06:50:27.793483Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-10T06:31:04.303077+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-10T06:31:04.303077+00:00","source":"crossref"},{"observed_at":"2026-08-10T06:30:57.382061+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2004.12993","last_updated":"2020-04-27T17:58:05Z","snapshot_observed_at":"2026-08-06T01:22:26.725556Z","submitted_at":"2020-04-27T17:58:05Z","title":"DeeBERT: Dynamic Early Exiting for Accelerating BERT Inference","version":1},"cited_work":{"arxiv_id":"2004.12993","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2004.12993","snapshot_observed_at":"2026-07-04T13:19:50.449132Z","title":"DeeBERT: Dynam ic Early Exiting for Accelerating BERT Inference","venue":null,"work_id":"1d9ce0aa-b85e-4fca-8bde-cc21a6e6e0f2","year":2020},"citing_paper":{"arxiv_id":"2605.02285","last_updated":"2026-05-04T07:19:48Z","snapshot_observed_at":"2026-08-02T21:49:54.214629Z","submitted_at":"2026-05-04T07:19:48Z","title":"Complexity Horizons of Compressed Models in Analog Circuit Analysis","version":1},"reference_index":22,"source":"pdf_text","source_observed_at":"2026-05-09T16:37:05.972866Z"},"links":{"cited_paper":"/paper/2004.12993","citing_paper":"/paper/2605.02285"},"observation_digest":"sha256:aa52bf175a51793a9be51d9e323954b4262de164c38a5ced7436061c89bd1d15","observation_id":"09a48e45-8264-4d8e-a009-b53da214255a","resolution":{"observed_at":"2026-05-11T16:26:10.480372Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-10T06:31:04.303077+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-10T06:31:04.303077+00:00","source":"crossref"},{"observed_at":"2026-08-10T06:30:57.382061+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2004.12993","last_updated":"2020-04-27T17:58:05Z","snapshot_observed_at":"2026-08-06T01:22:26.725556Z","submitted_at":"2020-04-27T17:58:05Z","title":"DeeBERT: Dynamic Early Exiting for Accelerating BERT Inference","version":1},"cited_work":{"arxiv_id":"2004.12993","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2004.12993","snapshot_observed_at":"2026-07-04T13:19:50.449132Z","title":"DeeBERT: Dynam ic Early Exiting for Accelerating BERT Inference","venue":null,"work_id":"1d9ce0aa-b85e-4fca-8bde-cc21a6e6e0f2","year":2020},"citing_paper":{"arxiv_id":"2605.05741","last_updated":"2026-05-07T06:32:57Z","snapshot_observed_at":"2026-07-06T23:18:17.333660Z","submitted_at":"2026-05-07T06:32:57Z","title":"HyperLens: Quantifying Cognitive Effort in LLMs with Fine-grained Confidence Trajectory","version":1},"reference_index":7,"source":"arxiv_source","source_observed_at":"2026-05-08T11:38:49.630171Z"},"links":{"cited_paper":"/paper/2004.12993","citing_paper":"/paper/2605.05741"},"observation_digest":"sha256:3bb8f0433a352a61733a26e3136d861cd11fc9644cbdd4efe794743153fc2ede","observation_id":"8f0bba89-c8c5-4349-b7e2-fd510daae1d1","resolution":{"observed_at":"2026-05-11T19:36:09.055857Z","resolver_source":"arxiv_id","status":"metadata_mismatch"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-10T06:31:04.303077+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-10T06:31:04.303077+00:00","source":"crossref"},{"observed_at":"2026-08-10T06:30:57.382061+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2004.12993","last_updated":"2020-04-27T17:58:05Z","snapshot_observed_at":"2026-08-06T01:22:26.725556Z","submitted_at":"2020-04-27T17:58:05Z","title":"DeeBERT: Dynamic Early Exiting for Accelerating BERT Inference","version":1},"cited_work":{"arxiv_id":"2004.12993","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2004.12993","snapshot_observed_at":"2026-07-04T13:19:50.449132Z","title":"DeeBERT: Dynam ic Early Exiting for Accelerating BERT Inference","venue":null,"work_id":"1d9ce0aa-b85e-4fca-8bde-cc21a6e6e0f2","year":2020},"citing_paper":{"arxiv_id":"2605.09165","last_updated":"2026-06-30T06:15:20Z","snapshot_observed_at":"2026-07-06T23:21:16.454273Z","submitted_at":"2026-05-09T20:58:18Z","title":"Sparse Layers are Critical to Scaling Looped Language Models","version":1},"reference_index":9,"source":"pdf_text","source_observed_at":"2026-05-12T04:19:12.696712Z"},"links":{"cited_paper":"/paper/2004.12993","citing_paper":"/paper/2605.09165"},"observation_digest":"sha256:cf1ad7479c1d3075e227a042b3e10260e51d56f4307b8eda2877e542d9a97b2d","observation_id":"6c79739b-5262-4d35-bb22-8bd0d2b8584c","resolution":{"observed_at":"2026-05-12T06:21:28.238775Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-10T06:31:04.303077+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-10T06:31:04.303077+00:00","source":"crossref"},{"observed_at":"2026-08-10T06:30:57.382061+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2004.12993","last_updated":"2020-04-27T17:58:05Z","snapshot_observed_at":"2026-08-06T01:22:26.725556Z","submitted_at":"2020-04-27T17:58:05Z","title":"DeeBERT: Dynamic Early Exiting for Accelerating BERT Inference","version":1},"cited_work":{"arxiv_id":"2004.12993","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2004.12993","snapshot_observed_at":"2026-07-04T13:19:50.449132Z","title":"DeeBERT: Dynam ic Early Exiting for Accelerating BERT Inference","venue":null,"work_id":"1d9ce0aa-b85e-4fca-8bde-cc21a6e6e0f2","year":2020},"citing_paper":{"arxiv_id":"2605.09165","last_updated":"2026-06-30T06:15:20Z","snapshot_observed_at":"2026-07-06T23:21:16.454273Z","submitted_at":"2026-05-09T20:58:18Z","title":"Sparse Layers are Critical to Scaling Looped Language Models","version":2},"reference_index":9,"source":"pdf_text","source_observed_at":"2026-07-01T07:26:37.559459Z"},"links":{"cited_paper":"/paper/2004.12993","citing_paper":"/paper/2605.09165"},"observation_digest":"sha256:8ea6649345c0024bb6a4052ca3e7e1b90174d22bd0af8528b32f22a3eb872289","observation_id":"12411a4e-616a-47d4-8169-a10b05fa6253","resolution":{"observed_at":"2026-07-01T07:35:29.162165Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-10T06:31:04.303077+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-10T06:31:04.303077+00:00","source":"crossref"},{"observed_at":"2026-08-10T06:30:57.382061+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2004.12993","last_updated":"2020-04-27T17:58:05Z","snapshot_observed_at":"2026-08-06T01:22:26.725556Z","submitted_at":"2020-04-27T17:58:05Z","title":"DeeBERT: Dynamic Early Exiting for Accelerating BERT Inference","version":1},"cited_work":{"arxiv_id":"2004.12993","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2004.12993","snapshot_observed_at":"2026-07-04T13:19:50.449132Z","title":"DeeBERT: Dynam ic Early Exiting for Accelerating BERT Inference","venue":null,"work_id":"1d9ce0aa-b85e-4fca-8bde-cc21a6e6e0f2","year":2020},"citing_paper":{"arxiv_id":"2605.10875","last_updated":"2026-05-11T17:27:15Z","snapshot_observed_at":"2026-07-06T23:22:47.781940Z","submitted_at":"2026-05-11T17:27:15Z","title":"Compute Where it Counts: Self Optimizing Language Models","version":1},"reference_index":22,"source":"pdf_text","source_observed_at":"2026-05-12T04:57:58.649358Z"},"links":{"cited_paper":"/paper/2004.12993","citing_paper":"/paper/2605.10875"},"observation_digest":"sha256:394fc9cdc1c8be7d5593c883c6e2648cd252897ab08ce9b236ae35dfef30a14f","observation_id":"ffde8642-6b5d-489b-bf19-e7a1170cc3be","resolution":{"observed_at":"2026-05-12T05:46:29.797977Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-10T06:31:04.303077+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-10T06:31:04.303077+00:00","source":"crossref"},{"observed_at":"2026-08-10T06:30:57.382061+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2004.12993","last_updated":"2020-04-27T17:58:05Z","snapshot_observed_at":"2026-08-06T01:22:26.725556Z","submitted_at":"2020-04-27T17:58:05Z","title":"DeeBERT: Dynamic Early Exiting for Accelerating BERT Inference","version":1},"cited_work":{"arxiv_id":"2004.12993","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2004.12993","snapshot_observed_at":"2026-07-04T13:19:50.449132Z","title":"DeeBERT: Dynam ic Early Exiting for Accelerating BERT Inference","venue":null,"work_id":"1d9ce0aa-b85e-4fca-8bde-cc21a6e6e0f2","year":2020},"citing_paper":{"arxiv_id":"2606.06574","last_updated":"2026-06-04T17:59:58Z","snapshot_observed_at":"2026-07-06T23:46:24.981618Z","submitted_at":"2026-06-04T17:59:58Z","title":"Skip a Layer or Loop It? Learning Program-of-Layers in LLMs","version":1},"reference_index":14,"source":"pdf_text","source_observed_at":"2026-06-28T01:56:34.435152Z"},"links":{"cited_paper":"/paper/2004.12993","citing_paper":"/paper/2606.06574"},"observation_digest":"sha256:f164947cd37d1feb4d1040c9d77d7d76785a30fba07066d5e0246715379f6383","observation_id":"cb36b86f-d44d-4757-b725-5863e1f413fb","resolution":{"observed_at":"2026-07-02T12:36:57.368620Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-10T06:31:04.303077+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-10T06:31:04.303077+00:00","source":"crossref"},{"observed_at":"2026-08-10T06:30:57.382061+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2004.12993","last_updated":"2020-04-27T17:58:05Z","snapshot_observed_at":"2026-08-06T01:22:26.725556Z","submitted_at":"2020-04-27T17:58:05Z","title":"DeeBERT: Dynamic Early Exiting for Accelerating BERT Inference","version":1},"cited_work":{"arxiv_id":"2004.12993","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2004.12993","snapshot_observed_at":"2026-07-04T13:19:50.449132Z","title":"DeeBERT: Dynam ic Early Exiting for Accelerating BERT Inference","venue":null,"work_id":"1d9ce0aa-b85e-4fca-8bde-cc21a6e6e0f2","year":2020},"citing_paper":{"arxiv_id":"2606.09937","last_updated":"2026-06-07T21:36:20Z","snapshot_observed_at":"2026-08-08T19:54:07.857285Z","submitted_at":"2026-06-07T21:36:20Z","title":"RKSC: Reasoning-Aware KV Cache Sharing and Confident Early Exit for Multi-Step LLM Inference","version":1},"reference_index":14,"source":"pdf_text","source_observed_at":"2026-06-27T18:29:16.153111Z"},"links":{"cited_paper":"/paper/2004.12993","citing_paper":"/paper/2606.09937"},"observation_digest":"sha256:29be9efd9e6963aa140ee20108b2b97558850d5eaaafa6aa22f5eb78ce4f5453","observation_id":"5719b8a9-4508-40d8-b385-326950dfdf37","resolution":{"observed_at":"2026-07-02T23:07:26.769760Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-10T06:31:04.303077+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-10T06:31:04.303077+00:00","source":"crossref"},{"observed_at":"2026-08-10T06:30:57.382061+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2004.12993","last_updated":"2020-04-27T17:58:05Z","snapshot_observed_at":"2026-08-06T01:22:26.725556Z","submitted_at":"2020-04-27T17:58:05Z","title":"DeeBERT: Dynamic Early Exiting for Accelerating BERT Inference","version":1},"cited_work":{"arxiv_id":"2004.12993","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2004.12993","snapshot_observed_at":"2026-07-04T13:19:50.449132Z","title":"DeeBERT: Dynam ic Early Exiting for Accelerating BERT Inference","venue":null,"work_id":"1d9ce0aa-b85e-4fca-8bde-cc21a6e6e0f2","year":2020},"citing_paper":{"arxiv_id":"2606.26538","last_updated":"2026-06-25T02:25:00Z","snapshot_observed_at":"2026-07-07T00:00:51.779266Z","submitted_at":"2026-06-25T02:25:00Z","title":"CascadeFormer: Depth-Tapered Transformers Motivated by Gradient Fan-in Asymmetry","version":1},"reference_index":40,"source":"arxiv_source","source_observed_at":"2026-06-26T05:22:26.818078Z"},"links":{"cited_paper":"/paper/2004.12993","citing_paper":"/paper/2606.26538"},"observation_digest":"sha256:a11c44989471b337127f40f45590cd5a015ae4696f533dc4ffd51094346d7e65","observation_id":"9ad77046-a3c2-477b-94ab-5334e38d8151","resolution":{"observed_at":"2026-07-04T13:19:50.450509Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-10T06:31:04.303077+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-10T06:31:04.303077+00:00","source":"crossref"},{"observed_at":"2026-08-10T06:30:57.382061+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2004.12993","last_updated":"2020-04-27T17:58:05Z","snapshot_observed_at":"2026-08-06T01:22:26.725556Z","submitted_at":"2020-04-27T17:58:05Z","title":"DeeBERT: Dynamic Early Exiting for Accelerating BERT Inference","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2004.12993","snapshot_observed_at":"2026-08-02T08:13:30.312749Z","title":"Liu Yang, Kangwook Lee, Robert Nowak, and Dimitris Papailiopoulos","venue":null,"work_id":null,"year":2004},"citing_paper":{"arxiv_id":"2607.20519","last_updated":"2026-07-08T02:54:03Z","snapshot_observed_at":"2026-08-06T23:41:10.107122Z","submitted_at":"2026-07-08T02:54:03Z","title":"Adaptive Depth in Looped Transformers: Diagnosing Learned Halting Gates and Trajectory Readouts","version":1},"reference_index":28,"source":"pdf_text","source_observed_at":"2026-08-02T08:13:30.312749Z"},"links":{"cited_paper":"/paper/2004.12993","citing_paper":"/paper/2607.20519"},"observation_digest":"sha256:0ceca8e59f8f117571513a1356038ba05afb2e309d51a01cbd659fdb97f28168","observation_id":"059860f5-77ec-4d59-8e36-817270d08a5f","resolution":{"observed_at":"2026-08-02T08:13:30.312749Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}}],"links":{"evidence":"/evidence","html":"/paper/2004.12993/citation-record","integrity":"/paper/2004.12993/integrity","json":"/paper/2004.12993/citation-record.json","paper":"/paper/2004.12993"},"outbound":[],"paper":{"arxiv_id":"2004.12993","last_updated":"2020-04-27T17:58:05Z","latest_version":1,"primary_category":"cs.CL","snapshot_observed_at":"2026-08-06T01:22:26.725556Z","submitted_at":"2020-04-27T17:58:05Z","title":"DeeBERT: Dynamic Early Exiting for Accelerating BERT Inference"},"reference_resolution":{"displayed":0,"state_counts":{"malformed_identifier":0,"metadata_mismatch":0,"parse_uncertain":0,"unresolved":0,"verified_exact":0,"verified_fuzzy":0},"total_outbound_references":0},"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":"2026-08-10T06:31:04.303077+00:00","source":"crossref"},{"observed_at":"2026-08-10T06:30:57.382061+00:00","source":"retraction_watch"}],"thesis":"As of 10 August 2026, this Paper Citation Record lists 0 of 0 outbound references and 23 inbound Pith citation observations for arXiv:2004.12993."}