{"as_of":"2026-08-21T15:11:00Z","caps":{"database_statements":6,"inbound":100,"outbound":100},"context_digest":"sha256:0d99a13c06431830fab0dd330d9d26f016cd2295a13167a78740e68981b84232","coverage":[{"denominator":25,"lane":"reference_resolution","note":"Typed states for the displayed outbound observations.","records_observed":25,"source":"paper_references, paper_reference_links","source_observed_at":"2026-05-19T17:55:35.764347Z","state":"measured"},{"denominator":26,"lane":"standing_notices","note":"One-hop event checks from named stored sources.","records_observed":26,"source":"scholarly_work_events, retraction_status_cache","source_observed_at":"2026-08-21T06:32:19.484+00:00","state":"measured"},{"denominator":1,"lane":"inbound_itemization","note":"Pith citing papers itemized under the disclosed page cap.","records_observed":1,"source":"paper_references, paper_reference_links","source_observed_at":"2026-08-01T08:38:52.168891Z","state":"measured"},{"denominator":1,"lane":"external_citation_measurements","note":"A source-named dated measurement, never combined with another source.","records_observed":0,"source":"cited_works","source_observed_at":null,"state":"measured"}],"external_citation_measurements":[],"inbound":[{"citation":{"cited_paper":{"arxiv_id":"2605.15208","last_updated":"2026-05-02T05:41:47Z","snapshot_observed_at":"2026-08-13T23:28:06.974131Z","submitted_at":"2026-05-02T05:41:47Z","title":"Quantization Undoes Alignment: Bias Emergence in Compressed LLMs Across Models and Precision Levels","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2605.15208","snapshot_observed_at":"2026-08-01T08:38:52.168891Z","title":"IEEE Cloud Summit , year=","venue":null,"work_id":null,"year":null},"citing_paper":{"arxiv_id":"2607.21063","last_updated":"2026-07-23T08:56:11Z","snapshot_observed_at":"2026-08-07T01:50:19.863208Z","submitted_at":"2026-07-23T08:56:11Z","title":"QuantiBias: Benchmarking Quantization-Induced Bias in LLMs","version":1},"reference_index":19,"source":"arxiv_source","source_observed_at":"2026-08-01T08:38:52.168891Z"},"links":{"cited_paper":"/paper/2605.15208","citing_paper":"/paper/2607.21063"},"observation_digest":"sha256:5af393d136bc5a0b5fb6b6a6739284a3f4b28bed9abaab76e9fcab925851e850","observation_id":"0d8c76db-04db-49c7-9c08-469a3e12f32f","resolution":{"observed_at":"2026-08-01T08:38:52.168891Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}}],"links":{"evidence":"/evidence","html":"/paper/2605.15208/citation-record","integrity":"/paper/2605.15208/integrity","json":"/paper/2605.15208/citation-record.json","paper":"/paper/2605.15208"},"outbound":[{"citation":{"cited_paper":{"arxiv_id":"2402.06196","last_updated":"2025-03-23T14:51:01Z","snapshot_observed_at":"2026-07-06T02:11:23.670680Z","submitted_at":"2024-02-09T05:37:09Z","title":"Large Language Models: A Survey","version":3},"cited_work":{"arxiv_id":"2402.06196","doi":"10.48550/arxiv.2402.06196","metadata_source":"pith","pith_arxiv_id":"2402.06196","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"Large Language Models: A Survey","venue":"cs.CL","work_id":"54e385fe-1786-48c3-8aa0-d727210eb50e","year":2024},"citing_paper":{"arxiv_id":"2605.15208","last_updated":"2026-05-02T05:41:47Z","snapshot_observed_at":"2026-08-13T23:28:06.974131Z","submitted_at":"2026-05-02T05:41:47Z","title":"Quantization Undoes Alignment: Bias Emergence in Compressed LLMs Across Models and Precision Levels","version":1},"reference_index":1,"source":"pdf_text","source_observed_at":"2026-05-19T17:55:35.764347Z"},"links":{"cited_paper":"/paper/2402.06196","citing_paper":"/paper/2605.15208"},"observation_digest":"sha256:ee78aff41b20fc117fe03acc91888f2b4172846f20c77687077e7979ef0c9b54","observation_id":"a9b94384-239e-45d5-805a-aacecfc74c42","resolution":{"observed_at":"2026-05-19T17:57:42.543982Z","resolver_source":"local_arxiv","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-21T06:32:19.484+00:00","source":"crossref"},{"observed_at":"2026-08-21T06:32:16.066871+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":"2026-06-05T21:23:00.469572Z","title":"A survey of post-training scaling in large language models","venue":null,"work_id":"398e275e-11e0-42ea-92cd-8105ad843d81","year":2025},"citing_paper":{"arxiv_id":"2605.15208","last_updated":"2026-05-02T05:41:47Z","snapshot_observed_at":"2026-08-13T23:28:06.974131Z","submitted_at":"2026-05-02T05:41:47Z","title":"Quantization Undoes Alignment: Bias Emergence in Compressed LLMs Across Models and Precision Levels","version":1},"reference_index":2,"source":"pdf_text","source_observed_at":"2026-05-19T17:55:35.764347Z"},"links":{"citing_paper":"/paper/2605.15208"},"observation_digest":"sha256:1ed383dfbe96e5864b03816c571c7a00f5884ff4b4e85e6368502aea92b7ff41","observation_id":"78c1b437-26bf-421f-a092-0808be762b66","resolution":{"observed_at":"2026-05-19T17:57:43.327045Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-21T06:32:19.484+00:00","source":"crossref"},{"observed_at":"2026-08-21T06:32:16.066871+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":"2026-06-05T21:23:00.469572Z","title":"LLMCBench: Benchmarking large language model com- pression for efficient deployment","venue":null,"work_id":"f9b50f1c-e835-4a44-8a53-649fe1e3e8f9","year":2024},"citing_paper":{"arxiv_id":"2605.15208","last_updated":"2026-05-02T05:41:47Z","snapshot_observed_at":"2026-08-13T23:28:06.974131Z","submitted_at":"2026-05-02T05:41:47Z","title":"Quantization Undoes Alignment: Bias Emergence in Compressed LLMs Across Models and Precision Levels","version":1},"reference_index":3,"source":"pdf_text","source_observed_at":"2026-05-19T17:55:35.764347Z"},"links":{"citing_paper":"/paper/2605.15208"},"observation_digest":"sha256:59ddb82c2603ee6fd24dc036d350fc2d7d9f9f234c345716af50cc0484bb3540","observation_id":"81f1a329-29f5-46a6-bae7-dd633ef67336","resolution":{"observed_at":"2026-05-19T17:57:43.324758Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-21T06:32:19.484+00:00","source":"crossref"},{"observed_at":"2026-08-21T06:32:16.066871+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":"2026-06-05T21:23:00.469572Z","title":"A survey of model compression techniques: Past, present, and future","venue":null,"work_id":"b8cbcbd5-4c00-490a-922b-ace6c7cfeee9","year":2025},"citing_paper":{"arxiv_id":"2605.15208","last_updated":"2026-05-02T05:41:47Z","snapshot_observed_at":"2026-08-13T23:28:06.974131Z","submitted_at":"2026-05-02T05:41:47Z","title":"Quantization Undoes Alignment: Bias Emergence in Compressed LLMs Across Models and Precision Levels","version":1},"reference_index":4,"source":"pdf_text","source_observed_at":"2026-05-19T17:55:35.764347Z"},"links":{"citing_paper":"/paper/2605.15208"},"observation_digest":"sha256:4366f7f53de300ef6e2e253f0fb347e253fabc885b18feae39f2d60392b95821","observation_id":"985048e7-75dc-4726-879f-8a060cc018ef","resolution":{"observed_at":"2026-05-19T17:57:43.316046Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-21T06:32:19.484+00:00","source":"crossref"},{"observed_at":"2026-08-21T06:32:16.066871+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2311.05232","last_updated":"2024-11-19T12:42:45Z","snapshot_observed_at":"2026-07-06T16:45:07.733095Z","submitted_at":"2023-11-09T09:25:37Z","title":"A Survey on Hallucination in Large Language Models: Principles, Taxonomy, Challenges, and Open Questions","version":2},"cited_work":{"arxiv_id":"2311.05232","doi":"10.48550/arxiv.2311.05232","metadata_source":"pith","pith_arxiv_id":"2311.05232","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"A Survey on Hallucination in Large Language Models: Principles, Taxonomy, Challenges, and Open Questions","venue":"cs.CL","work_id":"6389ea5a-8e37-498f-9bb2-a238ae4b3b9d","year":2023},"citing_paper":{"arxiv_id":"2605.15208","last_updated":"2026-05-02T05:41:47Z","snapshot_observed_at":"2026-08-13T23:28:06.974131Z","submitted_at":"2026-05-02T05:41:47Z","title":"Quantization Undoes Alignment: Bias Emergence in Compressed LLMs Across Models and Precision Levels","version":1},"reference_index":5,"source":"pdf_text","source_observed_at":"2026-05-19T17:55:35.764347Z"},"links":{"cited_paper":"/paper/2311.05232","citing_paper":"/paper/2605.15208"},"observation_digest":"sha256:1404eb550fced107add3c237617078b4f5f24ad859ff6b5f2822eb8f445179fe","observation_id":"471d4443-7597-4bec-933d-5325bf0677e7","resolution":{"observed_at":"2026-05-19T17:57:42.521114Z","resolver_source":"local_arxiv","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-21T06:32:19.484+00:00","source":"crossref"},{"observed_at":"2026-08-04T01:08:06.736062+00:00","source":"crossref_status_cache"},{"observed_at":"2026-08-04T01:08:06.736062+00:00","source":"openalex_status_cache"},{"observed_at":"2026-08-21T06:32:16.066871+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":"2504.1236","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-06-05T21:23:00.469572Z","title":"A survey on hallucination in large language and foundation models","venue":null,"work_id":"66bbbfaf-a566-4fcc-bce9-a983923eb02c","year":2025},"citing_paper":{"arxiv_id":"2605.15208","last_updated":"2026-05-02T05:41:47Z","snapshot_observed_at":"2026-08-13T23:28:06.974131Z","submitted_at":"2026-05-02T05:41:47Z","title":"Quantization Undoes Alignment: Bias Emergence in Compressed LLMs Across Models and Precision Levels","version":1},"reference_index":6,"source":"pdf_text","source_observed_at":"2026-05-19T17:55:35.764347Z"},"links":{"citing_paper":"/paper/2605.15208"},"observation_digest":"sha256:b70bd328be8b76b8312b08619550bf0555c7d119662a6bdb70253d1cddfe39c1","observation_id":"4e063007-2832-4169-b019-205129e3a238","resolution":{"observed_at":"2026-05-19T17:57:42.515664Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-21T06:32:19.484+00:00","source":"crossref"},{"observed_at":"2026-08-21T06:32:16.066871+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2502.19545","last_updated":"2025-07-21T15:24:26Z","snapshot_observed_at":"2026-08-18T10:15:43.803444Z","submitted_at":"2025-02-26T20:34:58Z","title":"Winning Big with Small Models: Knowledge Distillation vs. Self-Training for Reducing Hallucination in Product QA Agents","version":2},"cited_work":{"arxiv_id":"2502.19545","doi":"10.48550/arxiv.2502.19545","metadata_source":"arxiv_reference","pith_arxiv_id":"2502.19545","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"Winning big with small models: Knowledge distil- lation vs. self-training for reducing hallucination in QA agents","venue":"ArXiv.org","work_id":"00407250-2d44-4322-a156-f6eaaf49979f","year":2025},"citing_paper":{"arxiv_id":"2605.15208","last_updated":"2026-05-02T05:41:47Z","snapshot_observed_at":"2026-08-13T23:28:06.974131Z","submitted_at":"2026-05-02T05:41:47Z","title":"Quantization Undoes Alignment: Bias Emergence in Compressed LLMs Across Models and Precision Levels","version":1},"reference_index":7,"source":"pdf_text","source_observed_at":"2026-05-19T17:55:35.764347Z"},"links":{"cited_paper":"/paper/2502.19545","citing_paper":"/paper/2605.15208"},"observation_digest":"sha256:59cc16aeb1e2eba8e8d57612fb9456c2ca7503c3f394fb9ce741c4a1e2489966","observation_id":"fc0a8228-3235-4b12-9449-eb8e668b21f0","resolution":{"observed_at":"2026-05-19T17:57:42.512107Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-21T06:32:19.484+00:00","source":"crossref"},{"observed_at":"2026-08-21T06:32:16.066871+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":"2026-06-05T21:23:00.469572Z","title":"Bias and fairness in large language models: A survey","venue":null,"work_id":"8e712c32-beb6-4ed2-82ac-f15c3170ca01","year":2024},"citing_paper":{"arxiv_id":"2605.15208","last_updated":"2026-05-02T05:41:47Z","snapshot_observed_at":"2026-08-13T23:28:06.974131Z","submitted_at":"2026-05-02T05:41:47Z","title":"Quantization Undoes Alignment: Bias Emergence in Compressed LLMs Across Models and Precision Levels","version":1},"reference_index":8,"source":"pdf_text","source_observed_at":"2026-05-19T17:55:35.764347Z"},"links":{"citing_paper":"/paper/2605.15208"},"observation_digest":"sha256:5e40598c24a79474afad8c4566e933087f390efa4ec956624b90acad36124016","observation_id":"143f2603-292c-4235-b830-b3c2b6487410","resolution":{"observed_at":"2026-05-19T17:57:43.307852Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-21T06:32:19.484+00:00","source":"crossref"},{"observed_at":"2026-08-21T06:32:16.066871+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":"2026-06-05T21:23:00.469572Z","title":"Beyond perplexity: Multi-dimensional safety evaluation of LLM compression","venue":null,"work_id":"9971c1fc-0bfe-47d4-9ee6-a11747c6a9ed","year":2024},"citing_paper":{"arxiv_id":"2605.15208","last_updated":"2026-05-02T05:41:47Z","snapshot_observed_at":"2026-08-13T23:28:06.974131Z","submitted_at":"2026-05-02T05:41:47Z","title":"Quantization Undoes Alignment: Bias Emergence in Compressed LLMs Across Models and Precision Levels","version":1},"reference_index":9,"source":"pdf_text","source_observed_at":"2026-05-19T17:55:35.764347Z"},"links":{"citing_paper":"/paper/2605.15208"},"observation_digest":"sha256:547bd5fb78f669a40698037a064036c9129abd7968b749cb2136bbf49ad0092e","observation_id":"8452ba2d-36bf-46ef-a48c-cf11b1776990","resolution":{"observed_at":"2026-05-19T17:57:43.305476Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-21T06:32:19.484+00:00","source":"crossref"},{"observed_at":"2026-08-21T06:32:16.066871+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2306.15261","last_updated":"2023-06-27T07:44:25Z","snapshot_observed_at":"2026-08-16T15:20:50.639516Z","submitted_at":"2023-06-27T07:44:25Z","title":"A Survey on Out-of-Distribution Evaluation of Neural NLP Models","version":1},"cited_work":{"arxiv_id":"2306.15261","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2306.15261","snapshot_observed_at":"2026-06-05T21:23:00.469572Z","title":"A survey on out-of-distribution evalu- ation of neural NLP models","venue":null,"work_id":"cd0aeed7-03cf-4719-8ac2-1c1609777ba5","year":2023},"citing_paper":{"arxiv_id":"2605.15208","last_updated":"2026-05-02T05:41:47Z","snapshot_observed_at":"2026-08-13T23:28:06.974131Z","submitted_at":"2026-05-02T05:41:47Z","title":"Quantization Undoes Alignment: Bias Emergence in Compressed LLMs Across Models and Precision Levels","version":1},"reference_index":10,"source":"pdf_text","source_observed_at":"2026-05-19T17:55:35.764347Z"},"links":{"cited_paper":"/paper/2306.15261","citing_paper":"/paper/2605.15208"},"observation_digest":"sha256:3c6e5285827af963005a3a7d82cd1689db1f5a9e61b1f6e5ee08e437f14ef67e","observation_id":"bf0e9f81-fb65-4ea5-a49e-78c95e3c145b","resolution":{"observed_at":"2026-05-19T17:57:42.539865Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-21T06:32:19.484+00:00","source":"crossref"},{"observed_at":"2026-08-21T06:32:16.066871+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2508.15361","last_updated":"2025-08-21T08:43:35Z","snapshot_observed_at":"2026-08-15T17:59:42.381369Z","submitted_at":"2025-08-21T08:43:35Z","title":"A Survey on Large Language Model Benchmarks","version":1},"cited_work":{"arxiv_id":"2508.15361","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2508.15361","snapshot_observed_at":"2026-07-04T05:09:36.772485Z","title":"A survey on large language model benchmarks, 2025a","venue":null,"work_id":"8d88765b-fed2-44c6-9298-d24da64f7f20","year":2025},"citing_paper":{"arxiv_id":"2605.15208","last_updated":"2026-05-02T05:41:47Z","snapshot_observed_at":"2026-08-13T23:28:06.974131Z","submitted_at":"2026-05-02T05:41:47Z","title":"Quantization Undoes Alignment: Bias Emergence in Compressed LLMs Across Models and Precision Levels","version":1},"reference_index":11,"source":"pdf_text","source_observed_at":"2026-05-19T17:55:35.764347Z"},"links":{"cited_paper":"/paper/2508.15361","citing_paper":"/paper/2605.15208"},"observation_digest":"sha256:a52b3ad3869b5d59bfd9d3d725ec9753c310d792da1265464a5c2cf29863b430","observation_id":"902288b6-666d-4eca-9fbe-18693aa5ac5a","resolution":{"observed_at":"2026-05-19T17:57:42.491127Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-21T06:32:19.484+00:00","source":"crossref"},{"observed_at":"2026-08-21T06:32:16.066871+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2211.03013","last_updated":"2022-11-06T02:59:27Z","snapshot_observed_at":"2026-08-17T12:36:04.464467Z","submitted_at":"2022-11-06T02:59:27Z","title":"Robust Lottery Tickets for Pre-trained Language Models","version":1},"cited_work":{"arxiv_id":"2211.03013","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2211.03013","snapshot_observed_at":"2026-06-05T21:23:00.469572Z","title":"Ro- bust lottery tickets for pre-trained language models","venue":null,"work_id":"aadcbae4-1882-40cd-99f8-f31bde8c8a8d","year":2022},"citing_paper":{"arxiv_id":"2605.15208","last_updated":"2026-05-02T05:41:47Z","snapshot_observed_at":"2026-08-13T23:28:06.974131Z","submitted_at":"2026-05-02T05:41:47Z","title":"Quantization Undoes Alignment: Bias Emergence in Compressed LLMs Across Models and Precision Levels","version":1},"reference_index":12,"source":"pdf_text","source_observed_at":"2026-05-19T17:55:35.764347Z"},"links":{"cited_paper":"/paper/2211.03013","citing_paper":"/paper/2605.15208"},"observation_digest":"sha256:2a522dea2665d72b21ccb03f42005d9674456f0a8e03e360f7885ff46cba6494","observation_id":"4a18eb4c-ea0b-49a1-bfe0-376588f60048","resolution":{"observed_at":"2026-05-19T17:57:42.487055Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-21T06:32:19.484+00:00","source":"crossref"},{"observed_at":"2026-08-21T06:32:16.066871+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":"2026-06-05T21:23:00.469572Z","title":"Understanding and over- coming the challenges of efficient transformer quantization","venue":null,"work_id":"c549edf3-56e7-4e1f-ad64-b77ee27eb96b","year":2021},"citing_paper":{"arxiv_id":"2605.15208","last_updated":"2026-05-02T05:41:47Z","snapshot_observed_at":"2026-08-13T23:28:06.974131Z","submitted_at":"2026-05-02T05:41:47Z","title":"Quantization Undoes Alignment: Bias Emergence in Compressed LLMs Across Models and Precision Levels","version":1},"reference_index":13,"source":"pdf_text","source_observed_at":"2026-05-19T17:55:35.764347Z"},"links":{"citing_paper":"/paper/2605.15208"},"observation_digest":"sha256:0cf3ae9c2a836b9cac6c15b84aff0182ac8eeab12bed3f104f4f21bd1ff62d2e","observation_id":"a0aab113-6fbe-4e58-b677-32aa145baf5d","resolution":{"observed_at":"2026-05-19T17:57:43.322397Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-21T06:32:19.484+00:00","source":"crossref"},{"observed_at":"2026-08-21T06:32:16.066871+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":"2026-06-05T21:23:00.469572Z","title":"The compression techniques applied on deep learning model","venue":null,"work_id":"939b1041-52f7-4826-884e-73d6a06b319d","year":2022},"citing_paper":{"arxiv_id":"2605.15208","last_updated":"2026-05-02T05:41:47Z","snapshot_observed_at":"2026-08-13T23:28:06.974131Z","submitted_at":"2026-05-02T05:41:47Z","title":"Quantization Undoes Alignment: Bias Emergence in Compressed LLMs Across Models and Precision Levels","version":1},"reference_index":14,"source":"pdf_text","source_observed_at":"2026-05-19T17:55:35.764347Z"},"links":{"citing_paper":"/paper/2605.15208"},"observation_digest":"sha256:b5ac5eda3d07555c80d83711afcf17198478b82d7308f9ef4216410e72021565","observation_id":"2fd9e2ec-2442-4a61-afcd-dfac1a0d5907","resolution":{"observed_at":"2026-05-19T17:57:43.320302Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-21T06:32:19.484+00:00","source":"crossref"},{"observed_at":"2026-08-21T06:32:16.066871+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2401.03205","last_updated":"2024-01-06T12:40:45Z","snapshot_observed_at":"2026-08-19T05:45:32.734097Z","submitted_at":"2024-01-06T12:40:45Z","title":"The Dawn After the Dark: An Empirical Study on Factuality Hallucination in Large Language Models","version":1},"cited_work":{"arxiv_id":"2401.03205","doi":"10.48550/arxiv.2401.03205","metadata_source":"arxiv_reference","pith_arxiv_id":"2401.03205","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"X., Nie, J.-Y., and Wen, J.-R","venue":"arXiv (Cornell University)","work_id":"6d40e13d-4c37-4809-a89e-b6ad604e8d92","year":2024},"citing_paper":{"arxiv_id":"2605.15208","last_updated":"2026-05-02T05:41:47Z","snapshot_observed_at":"2026-08-13T23:28:06.974131Z","submitted_at":"2026-05-02T05:41:47Z","title":"Quantization Undoes Alignment: Bias Emergence in Compressed LLMs Across Models and Precision Levels","version":1},"reference_index":15,"source":"pdf_text","source_observed_at":"2026-05-19T17:55:35.764347Z"},"links":{"cited_paper":"/paper/2401.03205","citing_paper":"/paper/2605.15208"},"observation_digest":"sha256:d68501c7e7e50fe3b1812b5087346787a48bfe41ec1c51ed482755bb6730fa05","observation_id":"672bc7a9-9046-4b56-8729-0ff2fd426598","resolution":{"observed_at":"2026-05-19T17:57:42.499733Z","resolver_source":"arxiv_id","status":"metadata_mismatch"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-21T06:32:19.484+00:00","source":"crossref"},{"observed_at":"2026-08-21T06:32:16.066871+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":"2026-06-05T21:23:00.469572Z","title":"Understanding the effect of model compression on social bias in large language models","venue":null,"work_id":"e8c81fe3-bf52-4182-8e7e-59113b0f1c82","year":2023},"citing_paper":{"arxiv_id":"2605.15208","last_updated":"2026-05-02T05:41:47Z","snapshot_observed_at":"2026-08-13T23:28:06.974131Z","submitted_at":"2026-05-02T05:41:47Z","title":"Quantization Undoes Alignment: Bias Emergence in Compressed LLMs Across Models and Precision Levels","version":1},"reference_index":16,"source":"pdf_text","source_observed_at":"2026-05-19T17:55:35.764347Z"},"links":{"citing_paper":"/paper/2605.15208"},"observation_digest":"sha256:d4e60497383182121f9f3184fe76d19d3cd5f313603f40cfb8b39145dfad63d8","observation_id":"46b14417-7a05-4b87-b40d-8829e0af3514","resolution":{"observed_at":"2026-05-19T17:57:43.318211Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-21T06:32:19.484+00:00","source":"crossref"},{"observed_at":"2026-08-21T06:32:16.066871+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2407.03211","last_updated":"2024-10-12T17:26:41Z","snapshot_observed_at":"2026-08-16T13:37:20.485547Z","submitted_at":"2024-07-03T15:39:40Z","title":"How Does Quantization Affect Multilingual LLMs?","version":2},"cited_work":{"arxiv_id":"2407.03211","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2407.03211","snapshot_observed_at":"2026-07-02T01:56:27.980120Z","title":"How does quantization affect multilingual LLMs?","venue":null,"work_id":"0d979872-e4ef-4017-b6ef-7ef5ca15e6c5","year":2024},"citing_paper":{"arxiv_id":"2605.15208","last_updated":"2026-05-02T05:41:47Z","snapshot_observed_at":"2026-08-13T23:28:06.974131Z","submitted_at":"2026-05-02T05:41:47Z","title":"Quantization Undoes Alignment: Bias Emergence in Compressed LLMs Across Models and Precision Levels","version":1},"reference_index":17,"source":"pdf_text","source_observed_at":"2026-05-19T17:55:35.764347Z"},"links":{"cited_paper":"/paper/2407.03211","citing_paper":"/paper/2605.15208"},"observation_digest":"sha256:4345d5ab911b40bae3427adce6d23df4a78e2d98a122212fb0a71bd1f40ca481","observation_id":"f6704a4f-47df-4ec2-8dda-3f1118b95bef","resolution":{"observed_at":"2026-05-19T17:57:42.503592Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-21T06:32:19.484+00:00","source":"crossref"},{"observed_at":"2026-08-21T06:32:16.066871+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2502.15799","last_updated":"2025-06-29T19:20:45Z","snapshot_observed_at":"2026-08-16T12:57:10.993518Z","submitted_at":"2025-02-18T20:32:05Z","title":"Investigating the Impact of Quantization Methods on the Safety and Reliability of Large Language Models","version":2},"cited_work":{"arxiv_id":"2502.15799","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2502.15799","snapshot_observed_at":"2026-07-03T00:17:28.791413Z","title":"Investigating the impact of quantization methods on the safety and reliability of large language models","venue":null,"work_id":"08cabeeb-bd51-4582-8d2f-035fcf389d0c","year":2025},"citing_paper":{"arxiv_id":"2605.15208","last_updated":"2026-05-02T05:41:47Z","snapshot_observed_at":"2026-08-13T23:28:06.974131Z","submitted_at":"2026-05-02T05:41:47Z","title":"Quantization Undoes Alignment: Bias Emergence in Compressed LLMs Across Models and Precision Levels","version":1},"reference_index":18,"source":"pdf_text","source_observed_at":"2026-05-19T17:55:35.764347Z"},"links":{"cited_paper":"/paper/2502.15799","citing_paper":"/paper/2605.15208"},"observation_digest":"sha256:b14ac7b7fae10ed6139c7ecac3d38644965cd94d295afd33a590588640071743","observation_id":"eb2e1ace-dc33-4d63-8522-6058e770afea","resolution":{"observed_at":"2026-05-19T17:57:42.507878Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-21T06:32:19.484+00:00","source":"crossref"},{"observed_at":"2026-08-21T06:32:16.066871+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"1911.05248","last_updated":"2021-09-06T00:47:17Z","snapshot_observed_at":"2026-08-19T19:49:42.812040Z","submitted_at":"2019-11-13T02:02:19Z","title":"What Do Compressed Deep Neural Networks Forget?","version":3},"cited_work":{"arxiv_id":"1911.05248","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"1911.05248","snapshot_observed_at":"2026-07-04T07:09:38.117469Z","title":"What do compressed deep neural networks forget?arXiv preprint arXiv:1911.05248","venue":null,"work_id":"c9bc3423-e8b9-4934-b941-fe282832d49d","year":1911},"citing_paper":{"arxiv_id":"2605.15208","last_updated":"2026-05-02T05:41:47Z","snapshot_observed_at":"2026-08-13T23:28:06.974131Z","submitted_at":"2026-05-02T05:41:47Z","title":"Quantization Undoes Alignment: Bias Emergence in Compressed LLMs Across Models and Precision Levels","version":1},"reference_index":19,"source":"pdf_text","source_observed_at":"2026-05-19T17:55:35.764347Z"},"links":{"cited_paper":"/paper/1911.05248","citing_paper":"/paper/2605.15208"},"observation_digest":"sha256:6cc901afafcc4cd28b1c9b21cfabb83766c11848f78873ad3d97c6efa4341710","observation_id":"98905bc3-8204-4352-937a-2a3a3036ddc9","resolution":{"observed_at":"2026-05-19T17:57:42.535481Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-21T06:32:19.484+00:00","source":"crossref"},{"observed_at":"2026-08-21T06:32:16.066871+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":"2026-06-05T21:23:00.469572Z","title":"BBQ: A hand-built bias benchmark for question answering","venue":null,"work_id":"34efe4ac-ac23-492d-a509-bf62f9126d4a","year":2022},"citing_paper":{"arxiv_id":"2605.15208","last_updated":"2026-05-02T05:41:47Z","snapshot_observed_at":"2026-08-13T23:28:06.974131Z","submitted_at":"2026-05-02T05:41:47Z","title":"Quantization Undoes Alignment: Bias Emergence in Compressed LLMs Across Models and Precision Levels","version":1},"reference_index":20,"source":"pdf_text","source_observed_at":"2026-05-19T17:55:35.764347Z"},"links":{"citing_paper":"/paper/2605.15208"},"observation_digest":"sha256:96bf69f843323f9076ecb8311b0b40cbc833fe2a8271d35018908ead9fe1481e","observation_id":"e76988ec-6b84-452a-be1f-166ee7280d25","resolution":{"observed_at":"2026-05-19T17:57:43.313870Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-21T06:32:19.484+00:00","source":"crossref"},{"observed_at":"2026-08-21T06:32:16.066871+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":"2026-06-05T21:23:00.469572Z","title":"MLX: An array framework for Apple Silicon","venue":null,"work_id":"a6f422d2-ac57-4158-b0b9-5e26a934f57a","year":2023},"citing_paper":{"arxiv_id":"2605.15208","last_updated":"2026-05-02T05:41:47Z","snapshot_observed_at":"2026-08-13T23:28:06.974131Z","submitted_at":"2026-05-02T05:41:47Z","title":"Quantization Undoes Alignment: Bias Emergence in Compressed LLMs Across Models and Precision Levels","version":1},"reference_index":21,"source":"pdf_text","source_observed_at":"2026-05-19T17:55:35.764347Z"},"links":{"citing_paper":"/paper/2605.15208"},"observation_digest":"sha256:963dbc00ddc1f06da19fc8f8bdf0a786c93f2bbdacb5b6776b75562a501d76c3","observation_id":"5525bf5f-7525-4b71-ad42-af6c1f06599b","resolution":{"observed_at":"2026-05-19T17:57:43.311935Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-21T06:32:19.484+00:00","source":"crossref"},{"observed_at":"2026-08-21T06:32:16.066871+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":"2026-06-05T21:23:00.469572Z","title":"Cohen,Statistical Power Analysis for the Behavioral Sciences, 2nd ed","venue":null,"work_id":"367df29c-d07d-412e-be8b-ba0279300d85","year":1988},"citing_paper":{"arxiv_id":"2605.15208","last_updated":"2026-05-02T05:41:47Z","snapshot_observed_at":"2026-08-13T23:28:06.974131Z","submitted_at":"2026-05-02T05:41:47Z","title":"Quantization Undoes Alignment: Bias Emergence in Compressed LLMs Across Models and Precision Levels","version":1},"reference_index":22,"source":"pdf_text","source_observed_at":"2026-05-19T17:55:35.764347Z"},"links":{"citing_paper":"/paper/2605.15208"},"observation_digest":"sha256:0ef54ac3010094de7e665f171aef3ff47dacba4d80431269167129f82b078a8e","observation_id":"1dd28d74-8a3a-4d7d-b4e1-8954e0d54e1d","resolution":{"observed_at":"2026-05-19T17:57:43.309967Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-21T06:32:19.484+00:00","source":"crossref"},{"observed_at":"2026-08-21T06:32:16.066871+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":"2602.06181","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-07-01T19:16:00.127482Z","title":"Uncertainty drives social bias changes in quantized large language models","venue":null,"work_id":"b09d300e-237b-479a-a6cd-5a13b117dc25","year":2026},"citing_paper":{"arxiv_id":"2605.15208","last_updated":"2026-05-02T05:41:47Z","snapshot_observed_at":"2026-08-13T23:28:06.974131Z","submitted_at":"2026-05-02T05:41:47Z","title":"Quantization Undoes Alignment: Bias Emergence in Compressed LLMs Across Models and Precision Levels","version":1},"reference_index":23,"source":"pdf_text","source_observed_at":"2026-05-19T17:55:35.764347Z"},"links":{"citing_paper":"/paper/2605.15208"},"observation_digest":"sha256:12f7a65fe2ed4d42595576b7e89625c73840671470eccdc5bab5e3de5c05f71f","observation_id":"03be5546-2aa0-4ddc-83ae-6fd0cd91f904","resolution":{"observed_at":"2026-05-19T17:57:42.547977Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-21T06:32:19.484+00:00","source":"crossref"},{"observed_at":"2026-08-21T06:32:16.066871+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":"2511.07842","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-07-04T11:09:45.859028Z","title":"Alignment-aware quantization for LLM safety","venue":null,"work_id":"da16ed3b-a24a-42fc-b532-21fcf63ead88","year":2025},"citing_paper":{"arxiv_id":"2605.15208","last_updated":"2026-05-02T05:41:47Z","snapshot_observed_at":"2026-08-13T23:28:06.974131Z","submitted_at":"2026-05-02T05:41:47Z","title":"Quantization Undoes Alignment: Bias Emergence in Compressed LLMs Across Models and Precision Levels","version":1},"reference_index":24,"source":"pdf_text","source_observed_at":"2026-05-19T17:55:35.764347Z"},"links":{"citing_paper":"/paper/2605.15208"},"observation_digest":"sha256:045d0efefa4979cdb389712f5c5c43efd62210fd1a8e72b5666a7d6f5f119241","observation_id":"7763e137-dfaa-43d8-b682-d0740a956300","resolution":{"observed_at":"2026-05-19T17:57:42.495422Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-21T06:32:19.484+00:00","source":"crossref"},{"observed_at":"2026-08-21T06:32:16.066871+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2407.09141","last_updated":"2024-07-12T10:19:02Z","snapshot_observed_at":"2026-08-16T13:34:48.836282Z","submitted_at":"2024-07-12T10:19:02Z","title":"Accuracy is Not All You Need","version":1},"cited_work":{"arxiv_id":"2407.09141","doi":null,"metadata_source":"arxiv_reference","pith_arxiv_id":"2407.09141","snapshot_observed_at":"2026-07-01T16:25:49.793833Z","title":"Accuracy is not all you need","venue":null,"work_id":"c7adcd42-d86b-40d1-abd4-3fd2b2a1f497","year":2024},"citing_paper":{"arxiv_id":"2605.15208","last_updated":"2026-05-02T05:41:47Z","snapshot_observed_at":"2026-08-13T23:28:06.974131Z","submitted_at":"2026-05-02T05:41:47Z","title":"Quantization Undoes Alignment: Bias Emergence in Compressed LLMs Across Models and Precision Levels","version":1},"reference_index":25,"source":"pdf_text","source_observed_at":"2026-05-19T17:55:35.764347Z"},"links":{"cited_paper":"/paper/2407.09141","citing_paper":"/paper/2605.15208"},"observation_digest":"sha256:72e7d085e4a673d8e33e1ef0f4733f6f4d13c6d8f09c4e65f36c27a3096ffaf4","observation_id":"0f1e3dc1-6919-4ef8-9d56-4592b38885cf","resolution":{"observed_at":"2026-05-19T17:57:42.525569Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-21T06:32:19.484+00:00","source":"crossref"},{"observed_at":"2026-08-21T06:32:16.066871+00:00","source":"retraction_watch"}],"state":"measured"}}],"paper":{"arxiv_id":"2605.15208","last_updated":"2026-05-02T05:41:47Z","latest_version":1,"primary_category":"cs.LG","snapshot_observed_at":"2026-08-13T23:28:06.974131Z","submitted_at":"2026-05-02T05:41:47Z","title":"Quantization Undoes Alignment: Bias Emergence in Compressed LLMs Across Models and Precision Levels"},"reference_resolution":{"displayed":25,"state_counts":{"malformed_identifier":0,"metadata_mismatch":1,"parse_uncertain":0,"unresolved":0,"verified_exact":13,"verified_fuzzy":11},"total_outbound_references":25},"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":"2026-08-21T06:32:19.484+00:00","source":"crossref"},{"observed_at":"2026-08-21T06:32:16.066871+00:00","source":"retraction_watch"}],"thesis":"As of 21 August 2026, this Paper Citation Record lists 25 of 25 outbound references and 1 inbound Pith citation observation for arXiv:2605.15208."}