{"as_of":"2026-08-23T17:53:00Z","caps":{"database_statements":6,"inbound":100,"outbound":100},"context_digest":"sha256:13dfb66030e81826c491f10a44d02153d97252a587ca927c64e880b08c2e5635","coverage":[{"denominator":23,"lane":"reference_resolution","note":"Typed states for the displayed outbound observations.","records_observed":23,"source":"paper_references, paper_reference_links","source_observed_at":"2026-08-12T18:02:25.968729Z","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":"2026-08-23T06:30:58.430688+00:00","state":"measured"},{"denominator":16,"lane":"inbound_itemization","note":"Pith citing papers itemized under the disclosed page cap.","records_observed":16,"source":"paper_references, paper_reference_links","source_observed_at":"2026-08-14T04:14:46.739648Z","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":"2026-07-04T15:59:56.708231Z","state":"measured"}],"external_citation_measurements":[],"inbound":[{"citation":{"cited_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2411.17713","snapshot_observed_at":"2026-08-07T12:21:22.311989Z","title":"compe- tency","venue":null,"work_id":null,"year":2019},"citing_paper":{"arxiv_id":"2505.24621","last_updated":"2026-05-31T10:26:22Z","snapshot_observed_at":"2026-08-14T01:20:05.337302Z","submitted_at":"2025-05-30T14:12:07Z","title":"Benchmarking Large Language Models for Cryptanalysis and Side-Channel Vulnerabilities","version":3},"reference_index":3,"source":"pdf_text","source_observed_at":"2026-08-07T12:21:22.311989Z"},"links":{"cited_paper":"/paper/2411.17713","citing_paper":"/paper/2505.24621"},"observation_digest":"sha256:9c99be4b6909bf7e349d4f82ce27ea83d40fd8ebb1b695f3f0e3499635c072d8","observation_id":"65cd7426-e089-410f-8e19-f02b77191f2c","resolution":{"observed_at":"2026-08-07T12:21:22.311989Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2411.17713","snapshot_observed_at":"2026-08-06T16:12:48.091007Z","title":"Llama guard 3-1b-int4: Compact and efficient safeguard for human-ai conversations","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2507.14293","last_updated":"2025-07-18T18:06:27Z","snapshot_observed_at":"2026-08-20T02:31:40.859214Z","submitted_at":"2025-07-18T18:06:27Z","title":"WebGuard: Building a Generalizable Guardrail for Web Agents","version":1},"reference_index":14,"source":"pdf_text","source_observed_at":"2026-08-06T16:12:48.091007Z"},"links":{"cited_paper":"/paper/2411.17713","citing_paper":"/paper/2507.14293"},"observation_digest":"sha256:3c6006c5c0b294df37a8d9ac7cb44a265f408dcfa67b211a6b2b701e11109687","observation_id":"432a2faf-6660-47e1-bc12-ef20d58c848c","resolution":{"observed_at":"2026-08-06T16:12:48.091007Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2411.17713","snapshot_observed_at":"2026-08-06T13:05:33.120837Z","title":"Llama guard 3-1b-int4: Compact and efficient safeguard for human-ai conversations","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2507.21206","last_updated":"2025-07-28T17:58:12Z","snapshot_observed_at":"2026-08-14T13:44:31.305337Z","submitted_at":"2025-07-28T17:58:12Z","title":"Agentic Web: Weaving the Next Web with AI Agents","version":1},"reference_index":64,"source":"arxiv_source","source_observed_at":"2026-08-06T13:05:33.120837Z"},"links":{"cited_paper":"/paper/2411.17713","citing_paper":"/paper/2507.21206"},"observation_digest":"sha256:a3bc4c427c12f40f98c9305bca7ea1196e950cf6d78fb48a0c02fe10083a3884","observation_id":"236d537c-7a73-477f-bf10-2432a1dd882b","resolution":{"observed_at":"2026-08-06T13:05:33.120837Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"cited_work":{"arxiv_id":"2411.17713","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2411.17713","snapshot_observed_at":"2026-07-04T15:59:56.708231Z","title":"Llama guard 3-1b-int4: Compact and 9 efficient safeguard for human-ai conversations","venue":null,"work_id":"cbc9c821-66e2-4ab8-ab02-fb863905a462","year":2024},"citing_paper":{"arxiv_id":"2510.09689","last_updated":"2026-04-17T02:42:09Z","snapshot_observed_at":"2026-08-14T16:47:16.555613Z","submitted_at":"2025-10-09T09:44:14Z","title":"When Search Goes Wrong: Red-Teaming Web-Augmented Large Language Models","version":3},"reference_index":11,"source":"pdf_text","source_observed_at":"2026-05-18T09:29:14.842228Z"},"links":{"cited_paper":"/paper/2411.17713","citing_paper":"/paper/2510.09689"},"observation_digest":"sha256:bb91f661ffa805d246609dd3be19c33fce7c2314c3a9e84bb11e9bbfe7cf7f54","observation_id":"32d02be6-9e9b-4074-9c1f-c46998e03858","resolution":{"observed_at":"2026-05-18T09:31:11.609858Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-23T06:30:58.430688+00:00","source":"crossref"},{"observed_at":"2026-08-23T06:30:53.778098+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2411.17713","snapshot_observed_at":"2026-08-04T10:22:04.668209Z","title":"Fedorov, K","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2510.10271","last_updated":"2026-06-26T00:42:37Z","snapshot_observed_at":"2026-08-19T09:35:20.441231Z","submitted_at":"2025-10-11T16:14:56Z","title":"MetaBreak: Jailbreaking Online LLM Services via Special Token Manipulation","version":2},"reference_index":10,"source":"pdf_text","source_observed_at":"2026-08-04T10:22:04.668209Z"},"links":{"cited_paper":"/paper/2411.17713","citing_paper":"/paper/2510.10271"},"observation_digest":"sha256:ac149935d42e3c97190095af71aa1ce5ad4d789f15625be9871c9c6cd5c097f1","observation_id":"8a1409cd-e4e7-4cf8-b9a7-1e46bbcb643f","resolution":{"observed_at":"2026-08-04T10:22:04.668209Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2411.17713","snapshot_observed_at":"2026-08-04T09:47:23.420935Z","title":"Llama guard 3-1b-int4: Compact and efficient safeguard for human-ai conversations.arXiv preprint arXiv:2411.17713,","venue":null,"work_id":null,"year":null},"citing_paper":{"arxiv_id":"2510.13698","last_updated":"2026-07-14T06:18:50Z","snapshot_observed_at":"2026-08-15T11:10:34.711507Z","submitted_at":"2025-10-15T15:57:17Z","title":"Attention Misses Visual Risk: Risk-Adaptive Steering for Multimodal Safety Alignment","version":4},"reference_index":2025,"source":"pdf_text","source_observed_at":"2026-08-04T09:47:23.420935Z"},"links":{"cited_paper":"/paper/2411.17713","citing_paper":"/paper/2510.13698"},"observation_digest":"sha256:bb4d2eb275657496715abeabd65737bfdbca14bf83996167e37b8102fc847740","observation_id":"6970fe45-8b6d-4606-b3e2-32f5e9642be9","resolution":{"observed_at":"2026-08-04T09:47:23.420935Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"cited_work":{"arxiv_id":"2411.17713","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2411.17713","snapshot_observed_at":"2026-07-04T15:59:56.708231Z","title":"Llama guard 3-1b-int4: Compact and 9 efficient safeguard for human-ai conversations","venue":null,"work_id":"cbc9c821-66e2-4ab8-ab02-fb863905a462","year":2024},"citing_paper":{"arxiv_id":"2511.12710","last_updated":"2026-05-18T06:50:41Z","snapshot_observed_at":"2026-08-16T18:42:09.532465Z","submitted_at":"2025-11-16T17:52:07Z","title":"Evolve the Method, Not the Prompts: Evolutionary Synthesis of Jailbreak Attacks on LLMs","version":2},"reference_index":11,"source":"pdf_text","source_observed_at":"2026-05-21T18:58:53.183734Z"},"links":{"cited_paper":"/paper/2411.17713","citing_paper":"/paper/2511.12710"},"observation_digest":"sha256:64eeb3b6ad47fe4bbbe8ec56b3dabea822ab65857962595e7654ea26389865b5","observation_id":"04bc87d8-7e86-4d32-9e1b-ea13ea07b827","resolution":{"observed_at":"2026-05-21T19:00:30.359641Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-23T06:30:58.430688+00:00","source":"crossref"},{"observed_at":"2026-08-23T06:30:53.778098+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"cited_work":{"arxiv_id":"2411.17713","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2411.17713","snapshot_observed_at":"2026-07-04T15:59:56.708231Z","title":"Llama guard 3-1b-int4: Compact and 9 efficient safeguard for human-ai conversations","venue":null,"work_id":"cbc9c821-66e2-4ab8-ab02-fb863905a462","year":2024},"citing_paper":{"arxiv_id":"2603.15954","last_updated":"2026-04-27T22:25:48Z","snapshot_observed_at":"2026-07-06T22:49:19.323519Z","submitted_at":"2026-03-16T22:10:50Z","title":"MobileLLM-Flash: Latency-Guided On-Device LLM Design for Industry Scale Deployment","version":2},"reference_index":5,"source":"pdf_text","source_observed_at":"2026-05-15T09:43:09.767948Z"},"links":{"cited_paper":"/paper/2411.17713","citing_paper":"/paper/2603.15954"},"observation_digest":"sha256:9054eac8a33263cc6ae7aca8a713c875af3d6d3419bb21533a53a89c36b7fd6b","observation_id":"1c5544bc-177e-4899-85e1-df8d9b5e6168","resolution":{"observed_at":"2026-05-15T09:45:23.186472Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-23T06:30:58.430688+00:00","source":"crossref"},{"observed_at":"2026-08-23T06:30:53.778098+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"cited_work":{"arxiv_id":"2411.17713","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2411.17713","snapshot_observed_at":"2026-07-04T15:59:56.708231Z","title":"Llama guard 3-1b-int4: Compact and 9 efficient safeguard for human-ai conversations","venue":null,"work_id":"cbc9c821-66e2-4ab8-ab02-fb863905a462","year":2024},"citing_paper":{"arxiv_id":"2604.06552","last_updated":"2026-04-08T01:02:54Z","snapshot_observed_at":"2026-08-20T14:41:24.032728Z","submitted_at":"2026-04-08T01:02:54Z","title":"To Lie or Not to Lie? Investigating The Biased Spread of Global Lies by LLMs","version":1},"reference_index":14,"source":"arxiv_source","source_observed_at":"2026-05-10T18:41:54.081628Z"},"links":{"cited_paper":"/paper/2411.17713","citing_paper":"/paper/2604.06552"},"observation_digest":"sha256:f86750b69268f36ceb6624c9607540188ea5c1eb5c92728ac469a19f6a2afb94","observation_id":"f664cd62-6d4b-4024-85b2-e360d2b3c8b2","resolution":{"observed_at":"2026-05-11T00:05:50.810508Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-23T06:30:58.430688+00:00","source":"crossref"},{"observed_at":"2026-08-23T06:30:53.778098+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"cited_work":{"arxiv_id":"2411.17713","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2411.17713","snapshot_observed_at":"2026-07-04T15:59:56.708231Z","title":"Llama guard 3-1b-int4: Compact and 9 efficient safeguard for human-ai conversations","venue":null,"work_id":"cbc9c821-66e2-4ab8-ab02-fb863905a462","year":2024},"citing_paper":{"arxiv_id":"2605.17329","last_updated":"2026-05-17T08:35:38Z","snapshot_observed_at":"2026-08-15T02:24:24.397804Z","submitted_at":"2026-05-17T08:35:38Z","title":"LPG: Balancing Efficiency and Policy Reasoning in Latent Policy Guardrails","version":1},"reference_index":4,"source":"pdf_text","source_observed_at":"2026-05-19T23:42:58.135553Z"},"links":{"cited_paper":"/paper/2411.17713","citing_paper":"/paper/2605.17329"},"observation_digest":"sha256:acaa17f7a482c7d6f956184ae1a0b1f44e690bdc189c726713019147519d164d","observation_id":"405eafa7-6fa4-4b8b-b44a-6e942d79b480","resolution":{"observed_at":"2026-05-19T23:43:17.831171Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-23T06:30:58.430688+00:00","source":"crossref"},{"observed_at":"2026-08-23T06:30:53.778098+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"cited_work":{"arxiv_id":"2411.17713","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2411.17713","snapshot_observed_at":"2026-07-04T15:59:56.708231Z","title":"Llama guard 3-1b-int4: Compact and 9 efficient safeguard for human-ai conversations","venue":null,"work_id":"cbc9c821-66e2-4ab8-ab02-fb863905a462","year":2024},"citing_paper":{"arxiv_id":"2606.01513","last_updated":"2026-06-01T00:24:30Z","snapshot_observed_at":"2026-08-07T00:02:03.501606Z","submitted_at":"2026-06-01T00:24:30Z","title":"Compliance-Scored Best-of-N Guardrail Orchestration for Multimodal Document Generation in Payments Dispute Defense","version":1},"reference_index":16,"source":"pdf_text","source_observed_at":"2026-06-28T13:13:05.210457Z"},"links":{"cited_paper":"/paper/2411.17713","citing_paper":"/paper/2606.01513"},"observation_digest":"sha256:9660eb1234ea35ec21699111bc8cd9ea8a2a59c8cbd2f6a7c4530b9400236b93","observation_id":"6c4e3752-5a00-42d7-b875-cc056ba71925","resolution":{"observed_at":"2026-07-02T00:56:24.431995Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-23T06:30:58.430688+00:00","source":"crossref"},{"observed_at":"2026-08-23T06:30:53.778098+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"cited_work":{"arxiv_id":"2411.17713","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2411.17713","snapshot_observed_at":"2026-07-04T15:59:56.708231Z","title":"Llama guard 3-1b-int4: Compact and 9 efficient safeguard for human-ai conversations","venue":null,"work_id":"cbc9c821-66e2-4ab8-ab02-fb863905a462","year":2024},"citing_paper":{"arxiv_id":"2606.07968","last_updated":"2026-06-06T03:52:27Z","snapshot_observed_at":"2026-08-12T19:19:31.483970Z","submitted_at":"2026-06-06T03:52:27Z","title":"RecurGuard: Runtime Monitoring for Reasoning-Token Consumption Attacks","version":1},"reference_index":7,"source":"pdf_text","source_observed_at":"2026-06-27T19:45:30.671490Z"},"links":{"cited_paper":"/paper/2411.17713","citing_paper":"/paper/2606.07968"},"observation_digest":"sha256:f0cc6ae4106b0c2b60ead99075c130f056ee47625b84cae5b211bb1365161439","observation_id":"a609a6cc-21c8-4f50-86bf-b3e7784ea764","resolution":{"observed_at":"2026-07-02T21:27:24.556776Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-23T06:30:58.430688+00:00","source":"crossref"},{"observed_at":"2026-08-23T06:30:53.778098+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"cited_work":{"arxiv_id":"2411.17713","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2411.17713","snapshot_observed_at":"2026-07-04T15:59:56.708231Z","title":"Llama guard 3-1b-int4: Compact and 9 efficient safeguard for human-ai conversations","venue":null,"work_id":"cbc9c821-66e2-4ab8-ab02-fb863905a462","year":2024},"citing_paper":{"arxiv_id":"2606.09388","last_updated":"2026-06-08T12:03:51Z","snapshot_observed_at":"2026-08-13T14:47:40.331413Z","submitted_at":"2026-06-08T12:03:51Z","title":"Distilling Safe LLM Systems via Soft Prompts for On Device Settings","version":1},"reference_index":51,"source":"arxiv_source","source_observed_at":"2026-06-27T17:15:51.375580Z"},"links":{"cited_paper":"/paper/2411.17713","citing_paper":"/paper/2606.09388"},"observation_digest":"sha256:063bbf1311a03484e2b4523858cf121af93230f6748ed52f281bc4391f688eb5","observation_id":"d6766197-7316-4c61-abbb-fa4d168f2a9b","resolution":{"observed_at":"2026-07-03T00:27:29.253657Z","resolver_source":"arxiv_id","status":"metadata_mismatch"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-23T06:30:58.430688+00:00","source":"crossref"},{"observed_at":"2026-08-23T06:30:53.778098+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"cited_work":{"arxiv_id":"2411.17713","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2411.17713","snapshot_observed_at":"2026-07-04T15:59:56.708231Z","title":"Llama guard 3-1b-int4: Compact and 9 efficient safeguard for human-ai conversations","venue":null,"work_id":"cbc9c821-66e2-4ab8-ab02-fb863905a462","year":2024},"citing_paper":{"arxiv_id":"2606.26377","last_updated":"2026-06-24T21:00:39Z","snapshot_observed_at":"2026-08-13T23:20:47.685527Z","submitted_at":"2026-06-24T21:00:39Z","title":"Verifying Intent and Harm: A Unified Defense Against LLM-Generated Threats","version":1},"reference_index":10,"source":"pdf_text","source_observed_at":"2026-06-26T01:10:13.093255Z"},"links":{"cited_paper":"/paper/2411.17713","citing_paper":"/paper/2606.26377"},"observation_digest":"sha256:17846ecfd9c6e1ad6d5f2e10a11b61c87e898edb23903dc62ce2cd95f3ad02cd","observation_id":"dc87f997-f06f-4fe0-a3e7-a1192b8763a0","resolution":{"observed_at":"2026-07-04T15:59:56.709785Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-23T06:30:58.430688+00:00","source":"crossref"},{"observed_at":"2026-08-23T06:30:53.778098+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2411.17713","snapshot_observed_at":"2026-07-14T13:23:22.708604Z","title":"Gemma Team","venue":null,"work_id":null,"year":null},"citing_paper":{"arxiv_id":"2607.10226","last_updated":"2026-07-11T09:31:45Z","snapshot_observed_at":"2026-08-21T17:19:47.380915Z","submitted_at":"2026-07-11T09:31:45Z","title":"When Are Sparse Feature Interventions Actually Localized? Matched Evaluation for SAE-Based Safety Control","version":1},"reference_index":5,"source":"pdf_text","source_observed_at":"2026-07-14T13:23:22.708604Z"},"links":{"cited_paper":"/paper/2411.17713","citing_paper":"/paper/2607.10226"},"observation_digest":"sha256:3d2187a201cbb876e1e0f3afa66420b67e71617080b968baa22f314572067b7c","observation_id":"794b1101-7735-42a0-a3d8-a53e13c7148a","resolution":{"observed_at":"2026-07-14T13:23:22.708604Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2411.17713","snapshot_observed_at":"2026-08-14T04:14:46.739648Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2608.10279","last_updated":"2026-08-10T22:20:26Z","snapshot_observed_at":"2026-08-19T23:23:05.384986Z","submitted_at":"2026-08-10T22:20:26Z","title":"Withholding the Completing Chunk: Deterministic Pair-Completion Guardrails for Streaming LLM Output","version":1},"reference_index":15,"source":"pdf_text","source_observed_at":"2026-08-14T04:14:46.739648Z"},"links":{"cited_paper":"/paper/2411.17713","citing_paper":"/paper/2608.10279"},"observation_digest":"sha256:73491022e8e49071b9413c50849c4ab51b2c1f0d030cc255c66b3c93928c1c18","observation_id":"2d6eb7fe-b555-4fe0-9c43-77fc68c707d9","resolution":{"observed_at":"2026-08-14T04:14:46.739648Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}}],"links":{"evidence":"/evidence","html":"/paper/2411.17713/citation-record","integrity":"/paper/2411.17713/integrity","json":"/paper/2411.17713/citation-record.json","paper":"/paper/2411.17713"},"outbound":[{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-12T18:02:25.872059Z","title":"write newline","venue":null,"work_id":null,"year":null},"citing_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"reference_index":1,"source":"arxiv_source","source_observed_at":"2026-08-12T18:02:25.872059Z"},"links":{"citing_paper":"/paper/2411.17713"},"observation_digest":"sha256:fc2df70595e72dfd78379aab2fe9c4d9e4aa2d28f2803ae885a48227f632e8f5","observation_id":"ec53149a-042b-4281-9fc4-e7bc053d77e1","resolution":{"observed_at":"2026-08-12T18:02:25.872059Z","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":"2026-08-12T18:02:26.259965Z","title":"https://ai.meta.com/blog/llama-3-2-connect-2024-vision-edge-mobile-devices/","venue":null,"work_id":"53060199-38c8-4b50-81f2-24a9525a58e9","year":2024},"citing_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"reference_index":2,"source":"arxiv_source","source_observed_at":"2026-08-12T18:02:25.878295Z"},"links":{"citing_paper":"/paper/2411.17713"},"observation_digest":"sha256:d5a57386672dd873d61b0c0847c24a84693465838cb93918a74b6afc00bb06dc","observation_id":"7a8b78e5-f8bf-4fbe-8866-31841d6c45d8","resolution":{"observed_at":"2026-08-12T18:02:26.263883Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-23T06:30:58.430688+00:00","source":"crossref"},{"observed_at":"2026-08-23T06:30:53.778098+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2204.05862","last_updated":"2022-04-12T15:02:38Z","snapshot_observed_at":"2026-07-06T02:11:23.670680Z","submitted_at":"2022-04-12T15:02:38Z","title":"Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2204.05862","snapshot_observed_at":"2026-08-12T18:02:25.883378Z","title":"Training a helpful and harmless assistant with reinforcement learning from human feedback, 2022","venue":null,"work_id":null,"year":2022},"citing_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"reference_index":3,"source":"arxiv_source","source_observed_at":"2026-08-12T18:02:25.883378Z"},"links":{"cited_paper":"/paper/2204.05862","citing_paper":"/paper/2411.17713"},"observation_digest":"sha256:6cd4349d74ea86d0149439c7bacc7681c5b36829ae55e1cf198c6daac7b6954e","observation_id":"ca20e436-acf1-4005-aaab-8b2b845c16c5","resolution":{"observed_at":"2026-08-12T18:02:25.883378Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"1308.3432","last_updated":"2013-08-15T15:19:34Z","snapshot_observed_at":"2026-08-14T04:51:04.817737Z","submitted_at":"2013-08-15T15:19:34Z","title":"Estimating or Propagating Gradients Through Stochastic Neurons for Conditional Computation","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"1308.3432","snapshot_observed_at":"2026-08-12T18:02:25.888041Z","title":"Estimating or propagating gradients through stochastic neurons for conditional computation","venue":null,"work_id":null,"year":2013},"citing_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"reference_index":4,"source":"arxiv_source","source_observed_at":"2026-08-12T18:02:25.888041Z"},"links":{"cited_paper":"/paper/1308.3432","citing_paper":"/paper/2411.17713"},"observation_digest":"sha256:f636d6eaca22314863e8c1398425df0f44bcd6b30e83009d767da6ffb3d1fdb4","observation_id":"438e7723-d9be-490f-9d23-9ab6ba9e7048","resolution":{"observed_at":"2026-08-12T18:02:25.888041Z","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":"2026-08-12T18:02:26.248194Z","title":"Privileged bases in the transformer residual stream, 2023","venue":null,"work_id":"44d1c42e-c7d2-432b-aca1-62887c995229","year":2023},"citing_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"reference_index":5,"source":"arxiv_source","source_observed_at":"2026-08-12T18:02:25.892812Z"},"links":{"citing_paper":"/paper/2411.17713"},"observation_digest":"sha256:faded26e4b60f8315e7e25973ccff262277a40adda1cd81dc487c76c2aac6563","observation_id":"54075120-ec6d-4a68-9354-e9f890521335","resolution":{"observed_at":"2026-08-12T18:02:26.252030Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-23T06:30:58.430688+00:00","source":"crossref"},{"observed_at":"2026-08-23T06:30:53.778098+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":"2026-08-12T18:02:26.236032Z","title":"Executorch llama android demo app","venue":null,"work_id":"bff1f133-44b2-444b-9ac8-3c72d94e8ebf","year":2024},"citing_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"reference_index":6,"source":"arxiv_source","source_observed_at":"2026-08-12T18:02:25.897099Z"},"links":{"citing_paper":"/paper/2411.17713"},"observation_digest":"sha256:d7a81ea37b2122ea0446388de16645fe7e489c2a1085a028d9f379712b337ba4","observation_id":"9445110e-1c93-4aa3-b7b7-9f859188d44c","resolution":{"observed_at":"2026-08-12T18:02:26.240280Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-23T06:30:58.430688+00:00","source":"crossref"},{"observed_at":"2026-08-23T06:30:53.778098+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":"2026-08-12T18:02:26.223867Z","title":"Executorch llama ios demo app","venue":null,"work_id":"3c156a55-2154-4760-8f9d-4be2f5fa34dd","year":2024},"citing_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"reference_index":7,"source":"arxiv_source","source_observed_at":"2026-08-12T18:02:25.901076Z"},"links":{"citing_paper":"/paper/2411.17713"},"observation_digest":"sha256:bfee5aa63ae1cc6e601c1898276742eaa4e3bca43ccba2c7e7147309582c048d","observation_id":"77d8bacf-e03e-445e-b3b9-a4626d266ac5","resolution":{"observed_at":"2026-08-12T18:02:26.227918Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-23T06:30:58.430688+00:00","source":"crossref"},{"observed_at":"2026-08-23T06:30:53.778098+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"1503.02531","last_updated":"2015-03-09T15:44:49Z","snapshot_observed_at":"2026-08-16T18:00:58.008096Z","submitted_at":"2015-03-09T15:44:49Z","title":"Distilling the Knowledge in a Neural Network","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"1503.02531","snapshot_observed_at":"2026-08-12T18:02:25.905410Z","title":"Distilling the knowledge in a neural network","venue":null,"work_id":null,"year":2015},"citing_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"reference_index":8,"source":"arxiv_source","source_observed_at":"2026-08-12T18:02:25.905410Z"},"links":{"cited_paper":"/paper/1503.02531","citing_paper":"/paper/2411.17713"},"observation_digest":"sha256:c1d18100f759e43fa529697e9790c940460ef3721a27ae6040420bc9f258b851","observation_id":"09cd50c0-9d9a-4d2c-aa4c-15f9132dee53","resolution":{"observed_at":"2026-08-12T18:02:25.905410Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2312.06674","last_updated":"2023-12-07T19:40:50Z","snapshot_observed_at":"2026-08-14T15:42:19.849118Z","submitted_at":"2023-12-07T19:40:50Z","title":"Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2312.06674","snapshot_observed_at":"2026-08-12T18:02:25.909740Z","title":"Llama guard: Llm-based input-output safeguard for human-ai conversations, 2023","venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"reference_index":9,"source":"arxiv_source","source_observed_at":"2026-08-12T18:02:25.909740Z"},"links":{"cited_paper":"/paper/2312.06674","citing_paper":"/paper/2411.17713"},"observation_digest":"sha256:7f53e008c47b93610dc9ca8fb1eac88774ee1b50e65e29d2b5dad0dd1b9cce5c","observation_id":"dd29a1e1-b40a-4261-8d85-31afc142070a","resolution":{"observed_at":"2026-08-12T18:02:25.909740Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"1806.08342","last_updated":"2018-06-21T17:32:46Z","snapshot_observed_at":"2026-08-15T01:28:14.336267Z","submitted_at":"2018-06-21T17:32:46Z","title":"Quantizing deep convolutional networks for efficient inference: A whitepaper","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"1806.08342","snapshot_observed_at":"2026-08-12T18:02:25.913945Z","title":"Quantizing deep convolutional networks for efficient inference: A whitepaper","venue":null,"work_id":null,"year":2018},"citing_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"reference_index":10,"source":"arxiv_source","source_observed_at":"2026-08-12T18:02:25.913945Z"},"links":{"cited_paper":"/paper/1806.08342","citing_paper":"/paper/2411.17713"},"observation_digest":"sha256:f5f48f57adc577841aaabddcf22dbbfd68a62afc21eb8a06ba76f12d764b88ec","observation_id":"7ff4dc5e-4108-428e-9d43-6b9eb8138048","resolution":{"observed_at":"2026-08-12T18:02:25.913945Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2305.17888","last_updated":"2023-05-29T05:22:11Z","snapshot_observed_at":"2026-08-21T22:58:47.983191Z","submitted_at":"2023-05-29T05:22:11Z","title":"LLM-QAT: Data-Free Quantization Aware Training for Large Language Models","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2305.17888","snapshot_observed_at":"2026-08-12T18:02:25.918193Z","title":"Llm-qat: Data-free quantization aware training for large language models","venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"reference_index":11,"source":"arxiv_source","source_observed_at":"2026-08-12T18:02:25.918193Z"},"links":{"cited_paper":"/paper/2305.17888","citing_paper":"/paper/2411.17713"},"observation_digest":"sha256:2da275082feccf1a2662a5c56b96b009e69606857fda7dd660b816bcff997888","observation_id":"278114ff-0e6f-4768-99c9-7208e4e2f2fc","resolution":{"observed_at":"2026-08-12T18:02:25.918193Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2407.21783","last_updated":"2024-11-23T23:27:33Z","snapshot_observed_at":"2026-08-13T17:20:44.002518Z","submitted_at":"2024-07-31T17:54:27Z","title":"The Llama 3 Herd of Models","version":3},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2407.21783","snapshot_observed_at":"2026-08-12T18:02:25.922727Z","title":"The llama 3 herd of models, 2024 a","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"reference_index":12,"source":"arxiv_source","source_observed_at":"2026-08-12T18:02:25.922727Z"},"links":{"cited_paper":"/paper/2407.21783","citing_paper":"/paper/2411.17713"},"observation_digest":"sha256:3148b4791826216402cc24e736cf9b6c177d987b24889b64369ee176e1524435","observation_id":"68a2ad6a-5268-48b3-9dc9-d75265d818ed","resolution":{"observed_at":"2026-08-12T18:02:25.922727Z","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":"2026-08-12T18:02:26.211630Z","title":"Meta llama guard 2","venue":null,"work_id":"60b11408-39b8-485c-b29b-997d0454dbdb","year":2024},"citing_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"reference_index":13,"source":"arxiv_source","source_observed_at":"2026-08-12T18:02:25.926933Z"},"links":{"citing_paper":"/paper/2411.17713"},"observation_digest":"sha256:f5f93089bb664400ffee7abad2f2eb92988d3a9c4877b4ffb8204fdc9ea173b2","observation_id":"225443d4-243a-48d2-afaf-e7e19b24354b","resolution":{"observed_at":"2026-08-12T18:02:26.215623Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-23T06:30:58.430688+00:00","source":"crossref"},{"observed_at":"2026-08-23T06:30:53.778098+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":"2026-08-12T18:02:26.198485Z","title":"The llama 3 family of models","venue":null,"work_id":"5e9358ba-b250-4318-8887-4faa15f39dee","year":2024},"citing_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"reference_index":14,"source":"arxiv_source","source_observed_at":"2026-08-12T18:02:25.930256Z"},"links":{"citing_paper":"/paper/2411.17713"},"observation_digest":"sha256:c591a61998f2f5ed3235d080a5954f27e5227eec5e1044b25d096c7625885c6e","observation_id":"e5423069-89b1-43fc-8182-49538f865f9c","resolution":{"observed_at":"2026-08-12T18:02:26.203120Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-23T06:30:58.430688+00:00","source":"crossref"},{"observed_at":"2026-08-23T06:30:53.778098+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2403.03853","last_updated":"2024-10-11T09:43:32Z","snapshot_observed_at":"2026-08-16T14:12:14.696997Z","submitted_at":"2024-03-06T17:04:18Z","title":"ShortGPT: Layers in Large Language Models are More Redundant Than You Expect","version":3},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.03853","snapshot_observed_at":"2026-08-12T18:02:25.933651Z","title":"Shortgpt: Layers in large language models are more redundant than you expect, 2024","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"reference_index":15,"source":"arxiv_source","source_observed_at":"2026-08-12T18:02:25.933651Z"},"links":{"cited_paper":"/paper/2403.03853","citing_paper":"/paper/2411.17713"},"observation_digest":"sha256:d186162341886508e774405aa2905ca2cdad048acb73ec3c49c9d968a248a839","observation_id":"019f4aa4-837d-4455-8a49-b47f4620fc7d","resolution":{"observed_at":"2026-08-12T18:02:25.933651Z","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":"2026-08-12T18:02:26.184214Z","title":"Announcing mlcommons ai safety v0.5 proof of concept","venue":null,"work_id":"a6de482d-7488-4c47-b2ed-56932277d791","year":2024},"citing_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"reference_index":16,"source":"arxiv_source","source_observed_at":"2026-08-12T18:02:25.937877Z"},"links":{"citing_paper":"/paper/2411.17713"},"observation_digest":"sha256:403cd9e5624cec9edcc3005763887316a207422f9ba97592ce690d6c04da954c","observation_id":"ed0a94ab-8932-490c-a35e-2083849da3a0","resolution":{"observed_at":"2026-08-12T18:02:26.188526Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-23T06:30:58.430688+00:00","source":"crossref"},{"observed_at":"2026-08-23T06:30:53.778098+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2407.14679","last_updated":"2024-11-04T17:36:38Z","snapshot_observed_at":"2026-08-16T13:32:35.013705Z","submitted_at":"2024-07-19T21:47:57Z","title":"Compact Language Models via Pruning and Knowledge Distillation","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2407.14679","snapshot_observed_at":"2026-08-12T18:02:25.941881Z","title":"Compact language models via pruning and knowledge distillation, 2024","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"reference_index":17,"source":"arxiv_source","source_observed_at":"2026-08-12T18:02:25.941881Z"},"links":{"cited_paper":"/paper/2407.14679","citing_paper":"/paper/2411.17713"},"observation_digest":"sha256:86ac211618d92342ca470ef92c7c53d61b56405ee8ecc187e2964564f41677d1","observation_id":"592cd71e-ea04-40b8-8434-410ee06276a9","resolution":{"observed_at":"2026-08-12T18:02:25.941881Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2106.08295","last_updated":"2021-06-15T17:12:42Z","snapshot_observed_at":"2026-08-19T05:42:08.537901Z","submitted_at":"2021-06-15T17:12:42Z","title":"A White Paper on Neural Network Quantization","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2106.08295","snapshot_observed_at":"2026-08-12T18:02:25.946798Z","title":"A white paper on neural network quantization","venue":null,"work_id":null,"year":2021},"citing_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"reference_index":18,"source":"arxiv_source","source_observed_at":"2026-08-12T18:02:25.946798Z"},"links":{"cited_paper":"/paper/2106.08295","citing_paper":"/paper/2411.17713"},"observation_digest":"sha256:b514fc870a9a73d02f537c81b0216e758872d5849699abf4bce6f0ed2928fca2","observation_id":"339b9a36-360d-4414-9f43-dc7e8becff32","resolution":{"observed_at":"2026-08-12T18:02:25.946798Z","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":"2026-08-12T18:02:26.170961Z","title":"Executorch runtime overview","venue":null,"work_id":"81a0cd02-69cd-41cd-a48d-9224b404fce9","year":2024},"citing_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"reference_index":19,"source":"arxiv_source","source_observed_at":"2026-08-12T18:02:25.951036Z"},"links":{"citing_paper":"/paper/2411.17713"},"observation_digest":"sha256:48a978a6f45295dc80ed2b5ae10d65525e63ff13f0bbe9c34c7d80eb0e79b00c","observation_id":"a01f716c-c72b-407d-a318-6c0182796c98","resolution":{"observed_at":"2026-08-12T18:02:26.175173Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-23T06:30:58.430688+00:00","source":"crossref"},{"observed_at":"2026-08-23T06:30:53.778098+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":"2026-08-12T18:02:26.156736Z","title":"Executorch xnnpack delegate","venue":null,"work_id":"0e225861-e071-4d74-ab54-4717a7bac537","year":2024},"citing_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"reference_index":20,"source":"arxiv_source","source_observed_at":"2026-08-12T18:02:25.955253Z"},"links":{"citing_paper":"/paper/2411.17713"},"observation_digest":"sha256:6bc893bb79e62b12e57bec730d7a1ec72a60280518e763922c86d516a69806f2","observation_id":"6b909aa4-560e-4847-b962-4047845e3cf2","resolution":{"observed_at":"2026-08-12T18:02:26.160658Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-23T06:30:58.430688+00:00","source":"crossref"},{"observed_at":"2026-08-23T06:30:53.778098+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2403.08295","last_updated":"2024-04-16T12:52:47Z","snapshot_observed_at":"2026-08-03T03:29:01.959523Z","submitted_at":"2024-03-13T06:59:16Z","title":"Gemma: Open Models Based on Gemini Research and Technology","version":4},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.08295","snapshot_observed_at":"2026-08-12T18:02:25.960573Z","title":"Gemma: Open models based on gemini research and technology","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"reference_index":21,"source":"arxiv_source","source_observed_at":"2026-08-12T18:02:25.960573Z"},"links":{"cited_paper":"/paper/2403.08295","citing_paper":"/paper/2411.17713"},"observation_digest":"sha256:db5af51b8bd06f5019da960f6681681d5a5ac8807f16ee6f081c7a97f8516c03","observation_id":"3d49b6ea-9b9e-44b6-91c4-4b23b66b2ce9","resolution":{"observed_at":"2026-08-12T18:02:25.960573Z","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":"2026-08-12T18:02:26.143892Z","title":"torchao: Pytorch native quantization and sparsity for training and inference, October 2024","venue":null,"work_id":"47cd217a-f2f4-4a11-8877-b03429d78f5d","year":2024},"citing_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"reference_index":22,"source":"arxiv_source","source_observed_at":"2026-08-12T18:02:25.964670Z"},"links":{"citing_paper":"/paper/2411.17713"},"observation_digest":"sha256:a5b6e2da3cee74bb0616ff006b59bce75628a40adb5756c2c2468d7140475e3b","observation_id":"0acf404a-4f28-43ea-9825-f2e5451decd2","resolution":{"observed_at":"2026-08-12T18:02:26.148387Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-23T06:30:58.430688+00:00","source":"crossref"},{"observed_at":"2026-08-23T06:30:53.778098+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":"2026-08-12T18:02:26.129422Z","title":"torchtune: Pytorch's finetuning library, April 2024","venue":null,"work_id":"15f40d83-0c91-4572-843e-505a64f589ee","year":2024},"citing_paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations","version":1},"reference_index":23,"source":"arxiv_source","source_observed_at":"2026-08-12T18:02:25.968729Z"},"links":{"citing_paper":"/paper/2411.17713"},"observation_digest":"sha256:6617467c98bcfeb71e21d9663043ba0834b60045e94c04163bad530993006d8f","observation_id":"504ffe1b-d309-4873-8879-5aebc0291fe9","resolution":{"observed_at":"2026-08-12T18:02:26.135721Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-23T06:30:58.430688+00:00","source":"crossref"},{"observed_at":"2026-08-23T06:30:53.778098+00:00","source":"retraction_watch"}],"state":"measured"}}],"paper":{"arxiv_id":"2411.17713","last_updated":"2024-11-18T21:42:17Z","latest_version":1,"primary_category":"cs.DC","snapshot_observed_at":"2026-08-22T04:00:16.282603Z","submitted_at":"2024-11-18T21:42:17Z","title":"Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations"},"reference_resolution":{"displayed":23,"state_counts":{"malformed_identifier":0,"metadata_mismatch":0,"parse_uncertain":0,"unresolved":12,"verified_exact":0,"verified_fuzzy":11},"total_outbound_references":23},"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":"2026-08-23T06:30:58.430688+00:00","source":"crossref"},{"observed_at":"2026-08-23T06:30:53.778098+00:00","source":"retraction_watch"}],"thesis":"As of 23 August 2026, this Paper Citation Record lists 23 of 23 outbound references and 16 inbound Pith citation observations for arXiv:2411.17713."}