{"as_of":"2026-08-08T23:25:00Z","caps":{"database_statements":6,"inbound":100,"outbound":100},"context_digest":"sha256:5ee7728f49607aab1c69cbb01d09872609799a74e56556788ec8c83d9b7ab292","coverage":[{"denominator":59,"lane":"reference_resolution","note":"Typed states for the displayed outbound observations.","records_observed":59,"source":"paper_references, paper_reference_links","source_observed_at":"2026-08-08T22:23:28.738176Z","state":"measured"},{"denominator":63,"lane":"standing_notices","note":"One-hop event checks from named stored sources.","records_observed":63,"source":"scholarly_work_events, retraction_status_cache","source_observed_at":"2026-08-08T06:32:00.761636+00:00","state":"measured"},{"denominator":4,"lane":"inbound_itemization","note":"Pith citing papers itemized under the disclosed page cap.","records_observed":4,"source":"paper_references, paper_reference_links","source_observed_at":"2026-08-07T14:33:48.011356Z","state":"measured"},{"denominator":1,"lane":"external_citation_measurements","note":"A source-named dated measurement, never combined with another source.","records_observed":1,"source":"arxiv_reference","source_observed_at":"2026-08-05T02:28:24.338817Z","state":"measured"}],"external_citation_measurements":[{"count":0,"observed_at":"2026-08-05T02:28:24.338817Z","source":"arxiv_reference"}],"inbound":[{"citation":{"cited_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2502.04556","snapshot_observed_at":"2026-08-07T14:33:48.011356Z","title":null,"venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2505.18503","last_updated":"2025-05-24T04:45:45Z","snapshot_observed_at":"2026-08-07T23:09:57.753840Z","submitted_at":"2025-05-24T04:45:45Z","title":"Focus on What Matters: Enhancing Medical Vision-Language Models with Automatic Attention Alignment Tuning","version":1},"reference_index":36,"source":"arxiv_source","source_observed_at":"2026-08-07T14:33:48.011356Z"},"links":{"cited_paper":"/paper/2502.04556","citing_paper":"/paper/2505.18503"},"observation_digest":"sha256:a167904e9161cae4ae7d4802e489f0241d53077984df1bee26be5cf74a6c4adb","observation_id":"cd4c73eb-1e49-4e00-918c-6c288c0a0a60","resolution":{"observed_at":"2026-08-07T14:33:48.011356Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"cited_work":{"arxiv_id":"2502.04556","doi":"10.48550/arxiv.2502.04556","metadata_source":"arxiv_reference","pith_arxiv_id":"2502.04556","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"Truthflow: Truthful llm generation via representation flow correction","venue":"ArXiv.org","work_id":"e2c456a6-c5f6-4e2f-bd27-f789b4fd6a61","year":2024},"citing_paper":{"arxiv_id":"2602.02280","last_updated":"2026-05-12T03:47:11Z","snapshot_observed_at":"2026-07-06T22:44:09.804048Z","submitted_at":"2026-02-02T16:20:51Z","title":"RACC: Representation-Aware Coverage Criteria for LLM Safety Testing","version":2},"reference_index":49,"source":"pdf_text","source_observed_at":"2026-05-16T08:12:55.296932Z"},"links":{"cited_paper":"/paper/2502.04556","citing_paper":"/paper/2602.02280"},"observation_digest":"sha256:ed38419f7c0d33f53ccc804e4a59ce7b17c8e7190a008103bc0b23b5828c28d9","observation_id":"9a47c79d-66b7-49ac-9c4f-2ead060a57a0","resolution":{"observed_at":"2026-05-16T08:17:36.552368Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-08T06:32:00.761636+00:00","source":"crossref"},{"observed_at":"2026-08-08T06:31:55.24221+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"cited_work":{"arxiv_id":"2502.04556","doi":"10.48550/arxiv.2502.04556","metadata_source":"arxiv_reference","pith_arxiv_id":"2502.04556","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"Truthflow: Truthful llm generation via representation flow correction","venue":"ArXiv.org","work_id":"e2c456a6-c5f6-4e2f-bd27-f789b4fd6a61","year":2024},"citing_paper":{"arxiv_id":"2605.03907","last_updated":"2026-05-05T15:59:42Z","snapshot_observed_at":"2026-08-02T13:39:36.450690Z","submitted_at":"2026-05-05T15:59:42Z","title":"Steer Like the LLM: Activation Steering that Mimics Prompting","version":1},"reference_index":3,"source":"arxiv_source","source_observed_at":"2026-05-07T16:20:10.078995Z"},"links":{"cited_paper":"/paper/2502.04556","citing_paper":"/paper/2605.03907"},"observation_digest":"sha256:f99b4d1649b5d5a3689df965851129205257e757f1aeba303818bfb9b18be829","observation_id":"a1686da3-46da-49c8-8909-707253a82eef","resolution":{"observed_at":"2026-05-09T01:59:34.588602Z","resolver_source":"arxiv_id","status":"metadata_mismatch"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-08T06:32:00.761636+00:00","source":"crossref"},{"observed_at":"2026-08-08T06:31:55.24221+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"cited_work":{"arxiv_id":"2502.04556","doi":"10.48550/arxiv.2502.04556","metadata_source":"arxiv_reference","pith_arxiv_id":"2502.04556","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"Truthflow: Truthful llm generation via representation flow correction","venue":"ArXiv.org","work_id":"e2c456a6-c5f6-4e2f-bd27-f789b4fd6a61","year":2024},"citing_paper":{"arxiv_id":"2606.03096","last_updated":"2026-06-11T03:04:55Z","snapshot_observed_at":"2026-08-04T02:32:31.439459Z","submitted_at":"2026-06-02T03:35:00Z","title":"Can Factual Opinions Be Edited (Manipulated) in Large Language Models?","version":2},"reference_index":2,"source":"pdf_text","source_observed_at":"2026-06-28T10:30:47.603024Z"},"links":{"cited_paper":"/paper/2502.04556","citing_paper":"/paper/2606.03096"},"observation_digest":"sha256:6a350d9f53ed7bf39a50d1b85ef8a764561fd333bafc6b673d5a28ae169fc720","observation_id":"3d32d3e8-4142-44f1-b257-9eae980a5405","resolution":{"observed_at":"2026-07-02T02:56:29.000039Z","resolver_source":"arxiv_id","status":"metadata_mismatch"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-08T06:32:00.761636+00:00","source":"crossref"},{"observed_at":"2026-08-08T06:31:55.24221+00:00","source":"retraction_watch"}],"state":"measured"}}],"links":{"evidence":"/evidence","html":"/paper/2502.04556/citation-record","integrity":"/paper/2502.04556/integrity","json":"/paper/2502.04556/citation-record.json","paper":"/paper/2502.04556"},"outbound":[{"citation":{"cited_paper":{"arxiv_id":"2303.08774","last_updated":"2024-03-04T06:01:33Z","snapshot_observed_at":"2026-08-07T07:30:12.213965Z","submitted_at":"2023-03-15T17:15:04Z","title":"GPT-4 Technical Report","version":6},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2303.08774","snapshot_observed_at":"2026-08-08T22:23:27.192931Z","title":"L., Almeida, D., Altenschmidt, J., Altman, S., Anadkat, S., et al","venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":1,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:27.192931Z"},"links":{"cited_paper":"/paper/2303.08774","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:5543df5f251897529cc9cab2c00e88988c55d5c1721622ab5adf9ad36be5e265","observation_id":"a2864384-b0f7-47c7-bbdf-2082d5fbefd4","resolution":{"observed_at":"2026-08-08T22:23:27.192931Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2012.13255","last_updated":"2020-12-22T07:42:30Z","snapshot_observed_at":"2026-08-08T07:37:46.709970Z","submitted_at":"2020-12-22T07:42:30Z","title":"Intrinsic Dimensionality Explains the Effectiveness of Language Model Fine-Tuning","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2012.13255","snapshot_observed_at":"2026-08-08T22:23:27.254012Z","title":"Intrinsic dimensionality explains the effectiveness of language model fine-tuning","venue":null,"work_id":null,"year":2012},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":2,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:27.254012Z"},"links":{"cited_paper":"/paper/2012.13255","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:e3d79f4b08ad1d4667338e17d773c913938bc90ccefe7e164410c7d855d28b93","observation_id":"3818e60a-cff0-4246-b839-04b28c61f2ec","resolution":{"observed_at":"2026-08-08T22:23:27.254012Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2304.13734","last_updated":"2023-10-17T09:34:30Z","snapshot_observed_at":"2026-08-01T19:59:13.992395Z","submitted_at":"2023-04-26T02:49:38Z","title":"The Internal State of an LLM Knows When It's Lying","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2304.13734","snapshot_observed_at":"2026-08-08T22:23:27.320617Z","title":"and Mitchell, T","venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":3,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:27.320617Z"},"links":{"cited_paper":"/paper/2304.13734","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:db8e13d148fc2c9f349f1947d3c308542419192339f40d254f94e597d3a36a98","observation_id":"e24c2f5f-b285-484e-8858-4cf600f8fc70","resolution":{"observed_at":"2026-08-08T22:23:27.320617Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2309.16609","last_updated":"2023-09-28T17:07:49Z","snapshot_observed_at":"2026-07-06T02:11:23.670680Z","submitted_at":"2023-09-28T17:07:49Z","title":"Qwen Technical Report","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2309.16609","snapshot_observed_at":"2026-08-08T22:23:27.350971Z","title":"Qwen technical report","venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":4,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:27.350971Z"},"links":{"cited_paper":"/paper/2309.16609","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:cff683145a821720c05aaa9e9f58bb59f6eba39cffb557636b4c27af25b54d3e","observation_id":"3cdc8482-1021-4ae2-b7b5-63c2a1112dcf","resolution":{"observed_at":"2026-08-08T22:23:27.350971Z","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":"2026-08-08T22:23:29.531817Z","title":"F., Liu, X., Jagadish, H., and Wang, L","venue":null,"work_id":"7ef3087c-cc96-4f21-9a96-69fcefe5dac5","year":2024},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":5,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:27.354733Z"},"links":{"citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:4e741238ec060a43dade905ce05067f06685c0061123066de66259629bbbae03","observation_id":"9c917423-b91d-4138-a46e-bc305100f03b","resolution":{"observed_at":"2026-08-08T22:23:29.561691Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-08T06:32:00.761636+00:00","source":"crossref"},{"observed_at":"2026-08-08T06:31:55.24221+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":"2026-08-08T22:23:29.484566Z","title":null,"venue":null,"work_id":"5211d1dd-ba97-42e1-ac85-7488ad75465d","year":2010},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":6,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:27.358845Z"},"links":{"citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:8c5c81f53ca0f679ecd1e52899ff1ea42c8a79d4baf4f21da2c080d25709ac69","observation_id":"3b003863-bb5e-474e-8730-6778ee43d89f","resolution":{"observed_at":"2026-08-08T22:23:29.487980Z","resolver_source":"raw_fallback","status":"unresolved"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-08T06:32:00.761636+00:00","source":"crossref"},{"observed_at":"2026-08-08T06:31:55.24221+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2406.02721","last_updated":"2024-10-12T08:30:33Z","snapshot_observed_at":"2026-07-06T18:25:31.706930Z","submitted_at":"2024-06-04T19:05:10Z","title":"Self-Control of LLM Behaviors by Compressing Suffix Gradient into Prefix Controller","version":3},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2406.02721","snapshot_observed_at":"2026-08-08T22:23:27.362479Z","title":"Self-control of llm behaviors by compressing suffix gradient into prefix controller","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":7,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:27.362479Z"},"links":{"cited_paper":"/paper/2406.02721","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:a1b15441ac366915115ddb9a6a382d9a9d700b552276ce6e90c720f76489ceb0","observation_id":"9dbca40c-cdbd-44d1-8860-f76e512d073c","resolution":{"observed_at":"2026-08-08T22:23:27.362479Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2406.00045","last_updated":"2024-07-29T18:19:35Z","snapshot_observed_at":"2026-08-07T01:32:56.141077Z","submitted_at":"2024-05-28T05:10:40Z","title":"Personalized Steering of Large Language Models: Versatile Steering Vectors Through Bi-directional Preference Optimization","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2406.00045","snapshot_observed_at":"2026-08-08T22:23:27.430340Z","title":"Personalized steering of large language models: Versatile steering vectors through bi-directional preference optimization","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":8,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:27.430340Z"},"links":{"cited_paper":"/paper/2406.00045","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:b2dddb962effc4eec42b1652fe1300e77a071d477a4f001b31aad1b0285f378e","observation_id":"0562b58a-3f5d-476b-87f7-72d7fd21ae64","resolution":{"observed_at":"2026-08-08T22:23:27.430340Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2307.15217","last_updated":"2023-09-11T17:25:24Z","snapshot_observed_at":"2026-08-03T19:11:09.671782Z","submitted_at":"2023-07-27T22:29:25Z","title":"Open Problems and Fundamental Limitations of Reinforcement Learning from Human Feedback","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2307.15217","snapshot_observed_at":"2026-08-08T22:23:27.507483Z","title":"K., Scheurer, J., Rando, J., Freedman, R., Korbak, T., Lindner, D., Freire, P., et al","venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":9,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:27.507483Z"},"links":{"cited_paper":"/paper/2307.15217","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:afff212f454ffc6d2feb4825e9522a516d2a154eae039d11e4f6842df532d7a4","observation_id":"8f12cd1d-928d-4dad-8eb7-ce13dc188a92","resolution":{"observed_at":"2026-08-08T22:23:27.507483Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2402.03744","last_updated":"2024-10-21T04:10:50Z","snapshot_observed_at":"2026-08-07T20:09:21.272663Z","submitted_at":"2024-02-06T06:23:12Z","title":"INSIDE: LLMs' Internal States Retain the Power of Hallucination Detection","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2402.03744","snapshot_observed_at":"2026-08-08T22:23:27.534094Z","title":"Inside: Llms' internal states retain the power of hallucination detection","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":10,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:27.534094Z"},"links":{"cited_paper":"/paper/2402.03744","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:8abf002a7de3df8fa61f8c2936f2501606789a063398c82660b4bb03de4a92c7","observation_id":"48e60b3d-9bdc-49ea-a7a6-ab95de8f0d4b","resolution":{"observed_at":"2026-08-08T22:23:27.534094Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2408.08769","last_updated":"2025-06-03T15:05:19Z","snapshot_observed_at":"2026-07-06T19:01:36.868186Z","submitted_at":"2024-08-16T14:23:59Z","title":"Lower Layers Matter: Alleviating Hallucination via Multi-Layer Fusion Contrastive Decoding with Truthfulness Refocused","version":2},"cited_work":{"arxiv_id":"2408.08769","doi":null,"metadata_source":"pith","pith_arxiv_id":"2408.08769","snapshot_observed_at":"2026-08-08T22:23:29.097209Z","title":"Lower Layers Matter: Alleviating Hallucination via Multi-Layer Fusion Contrastive Decoding with Truthfulness Refocused","venue":"cs.CL","work_id":"ae21c6c3-0f09-45b4-b029-af9d6f79a3b6","year":2024},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":11,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:27.538328Z"},"links":{"cited_paper":"/paper/2408.08769","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:958355333afd63eedeef3a0438082ff67acda1d68a902c5dbfedc601ee439321","observation_id":"fc1fb706-b76e-4937-9fcc-fd9c4400c00a","resolution":{"observed_at":"2026-08-08T22:23:29.149431Z","resolver_source":"local_arxiv","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-08T06:32:00.761636+00:00","source":"crossref"},{"observed_at":"2026-08-08T06:31:55.24221+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2403.01548","last_updated":"2024-03-12T09:49:28Z","snapshot_observed_at":"2026-07-06T17:38:51.653737Z","submitted_at":"2024-03-03T15:53:41Z","title":"In-Context Sharpness as Alerts: An Inner Representation Perspective for Hallucination Mitigation","version":3},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.01548","snapshot_observed_at":"2026-08-08T22:23:27.541432Z","title":"In-context sharpness as alerts: An inner representation perspective for hallucination mitigation","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":12,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:27.541432Z"},"links":{"cited_paper":"/paper/2403.01548","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:c0d9174348d3743de63f298a51dafdac27e472ff1925d73dcb23929714ed9ab4","observation_id":"df555452-e19f-42d0-b783-2653d6ac2cba","resolution":{"observed_at":"2026-08-08T22:23:27.541432Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2401.12292","last_updated":"2024-01-31T06:44:42Z","snapshot_observed_at":"2026-08-08T12:57:35.341622Z","submitted_at":"2024-01-22T19:00:08Z","title":"GRATH: Gradual Self-Truthifying for Large Language Models","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2401.12292","snapshot_observed_at":"2026-08-08T22:23:27.545341Z","title":"Grath: Gradual self-truthifying for large language models","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":13,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:27.545341Z"},"links":{"cited_paper":"/paper/2401.12292","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:68a047b2681adcadca76fac100048a6e561989c14c040ca7b9341192840cce61","observation_id":"0f2236e7-28e7-4254-bec9-28aa3f8d733f","resolution":{"observed_at":"2026-08-08T22:23:27.545341Z","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":"2026-08-08T22:23:29.473731Z","title":"Truth forest: Toward multi-scale truthfulness in large language models through intervention without tuning","venue":null,"work_id":"6c16cd04-2855-4428-a595-08fb33a5da71","year":2024},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":14,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:27.595829Z"},"links":{"citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:e31bba6514566450e79402e45371a53a786b56cfe6bd69d6ef4f2abdf7afe87e","observation_id":"201869d0-430e-4dd7-9084-999be4ef0c05","resolution":{"observed_at":"2026-08-08T22:23:29.477652Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-08T06:32:00.761636+00:00","source":"crossref"},{"observed_at":"2026-08-08T06:31:55.24221+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2309.03883","last_updated":"2024-03-11T02:01:09Z","snapshot_observed_at":"2026-08-06T11:43:44.726916Z","submitted_at":"2023-09-07T17:45:31Z","title":"DoLa: Decoding by Contrasting Layers Improves Factuality in Large Language Models","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2309.03883","snapshot_observed_at":"2026-08-08T22:23:27.687156Z","title":"Dola: Decoding by contrasting layers improves factuality in large language models","venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":15,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:27.687156Z"},"links":{"cited_paper":"/paper/2309.03883","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:01d3f7c65f7134181ce77b00f0a9f1b896a2c7835ce2e78352896934eb2eece9","observation_id":"3e289c78-b1fe-4411-ae9c-dc45a90603ba","resolution":{"observed_at":"2026-08-08T22:23:27.687156Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2409.17504","last_updated":"2024-09-26T03:22:09Z","snapshot_observed_at":"2026-08-03T18:50:25.972132Z","submitted_at":"2024-09-26T03:22:09Z","title":"HaloScope: Harnessing Unlabeled LLM Generations for Hallucination Detection","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2409.17504","snapshot_observed_at":"2026-08-08T22:23:27.744278Z","title":"Haloscope: Harnessing unlabeled llm generations for hallucination detection","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":16,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:27.744278Z"},"links":{"cited_paper":"/paper/2409.17504","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:cfc45613cbf0be774dcf7539a5e9e4280a66009dac57287c51bdfbfa95fed5dd","observation_id":"f68cfb77-be29-4af9-b752-46c146693ebb","resolution":{"observed_at":"2026-08-08T22:23:27.744278Z","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":"2026-07-06T18:55:11.576666Z","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":"2026-08-08T22:23:27.754085Z","title":"The llama 3 herd of models","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":17,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:27.754085Z"},"links":{"cited_paper":"/paper/2407.21783","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:6621c8756cfd59390f302057bce2fb757437f86e335c40bc7bde0a3e82910ea8","observation_id":"787cb7f5-6728-4149-9707-f5c3d3763c73","resolution":{"observed_at":"2026-08-08T22:23:27.754085Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-08T22:23:27.758535Z","title":"Scaling rectified flow transformers for high-resolution image synthesis","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":18,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:27.758535Z"},"links":{"citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:0cb01e52d74b736458bae38cf472b1c154085d4a43601d9470d0a40b5c2d053f","observation_id":"6c11c963-594f-4313-bb89-20419926711c","resolution":{"observed_at":"2026-08-08T22:23:27.758535Z","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":"2026-08-08T22:23:29.456820Z","title":"Institutionum calculi integralis, volume 4","venue":null,"work_id":"6091f300-f0d9-4b41-a962-776e52343390","year":null},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":19,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:27.762113Z"},"links":{"citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:72da993636654f87e1431d27d0fb341efec598385ba8fe878f1395ae0c187af8","observation_id":"83ad5505-3f12-4c00-bfce-1b79a568c449","resolution":{"observed_at":"2026-08-08T22:23:29.460372Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-08T06:32:00.761636+00:00","source":"crossref"},{"observed_at":"2026-08-08T06:31:55.24221+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2403.18680","last_updated":"2024-06-06T13:58:20Z","snapshot_observed_at":"2026-07-06T17:52:00.714976Z","submitted_at":"2024-03-27T15:22:16Z","title":"Non-Linear Inference Time Intervention: Improving LLM Truthfulness","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.18680","snapshot_observed_at":"2026-08-08T22:23:27.765808Z","title":"Nl-iti: Optimizing probing and intervention for improvement of iti method","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":20,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:27.765808Z"},"links":{"cited_paper":"/paper/2403.18680","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:f19bd77fde2bb097a851bd83f5279ca163668fea4fe83c9e414a5452a1c65a13","observation_id":"09f6ea34-236b-41c8-acbc-548facdc5fc8","resolution":{"observed_at":"2026-08-08T22:23:27.765808Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2406.11267","last_updated":"2024-06-17T07:16:07Z","snapshot_observed_at":"2026-08-07T22:55:07.093482Z","submitted_at":"2024-06-17T07:16:07Z","title":"Mitigating Large Language Model Hallucination with Faithful Finetuning","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2406.11267","snapshot_observed_at":"2026-08-08T22:23:27.819703Z","title":"Mitigating large language model hallucination with faithful finetuning","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":21,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:27.819703Z"},"links":{"cited_paper":"/paper/2406.11267","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:31f36712b4cbbbeeb95f4f5abee72db787339d0009f4738a5c35e974977b5400","observation_id":"a3af8d04-62d3-4be0-b553-0a72efb441f8","resolution":{"observed_at":"2026-08-08T22:23:27.819703Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-08T22:23:27.915938Z","title":"A survey on hallucination in large language models: Principles, taxonomy, challenges, and open questions","venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":22,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:27.915938Z"},"links":{"citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:acc5269d883fcea06088b6affe71e5df4ede2c2b44f021b859275f52db9040dc","observation_id":"68c342f0-6d71-40c0-bfc1-b1ef84d71f66","resolution":{"observed_at":"2026-08-08T22:23:27.915938Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-08T22:23:27.954263Z","title":"J., Madotto, A., and Fung, P","venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":23,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:27.954263Z"},"links":{"citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:2e5b46fa7a23e4fc27e2948841d7be86e38e5f278aa6b69f40b4341acf4f8427","observation_id":"a7f212f7-7f0f-4b11-a696-c8e4635fd91f","resolution":{"observed_at":"2026-08-08T22:23:27.954263Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2310.06825","last_updated":"2023-10-10T17:54:58Z","snapshot_observed_at":"2026-07-06T02:11:23.670680Z","submitted_at":"2023-10-10T17:54:58Z","title":"Mistral 7B","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2310.06825","snapshot_observed_at":"2026-08-08T22:23:27.962031Z","title":"Q., Sablayrolles, A., Mensch, A., Bamford, C., Chaplot, D","venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":24,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:27.962031Z"},"links":{"cited_paper":"/paper/2310.06825","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:8abe6ac4e516787f318f83ddd5d1218f09a83f3fd9c549bd01e98a9e81569a24","observation_id":"c86aba22-2696-4246-9a7f-c28467b2237c","resolution":{"observed_at":"2026-08-08T22:23:27.962031Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2404.07066","last_updated":"2025-02-04T23:34:30Z","snapshot_observed_at":"2026-07-06T17:58:21.138881Z","submitted_at":"2024-04-10T14:56:40Z","title":"Exploring Concept Depth: How Large Language Models Acquire Knowledge and Concept at Different Layers?","version":7},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2404.07066","snapshot_observed_at":"2026-08-08T22:23:27.966500Z","title":"Exploring concept depth: How large language models acquire knowledge at different layers? arXiv preprint arXiv:2404.07066, 2024","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":25,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:27.966500Z"},"links":{"cited_paper":"/paper/2404.07066","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:bcd75316c24f4c409134fb61ed80b14f1c0bc826065048fef041f59f374c6397","observation_id":"d4480f36-b2cd-493f-bf97-3b06d174a7db","resolution":{"observed_at":"2026-08-08T22:23:27.966500Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"1705.03551","last_updated":"2017-05-13T21:12:37Z","snapshot_observed_at":"2026-08-02T11:13:42.401488Z","submitted_at":"2017-05-09T21:35:07Z","title":"TriviaQA: A Large Scale Distantly Supervised Challenge Dataset for Reading Comprehension","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"1705.03551","snapshot_observed_at":"2026-08-08T22:23:27.970336Z","title":"S., and Zettlemoyer, L","venue":null,"work_id":null,"year":2017},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":26,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:27.970336Z"},"links":{"cited_paper":"/paper/1705.03551","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:e0fc92989b5ad4b3b068f2de2bd3a036bd77524fc28982921e2b7bba7c82be60","observation_id":"b3e1e005-5edf-4283-ae6e-3b368ca8b285","resolution":{"observed_at":"2026-08-08T22:23:27.970336Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2401.05930","last_updated":"2024-10-07T09:58:48Z","snapshot_observed_at":"2026-08-05T03:19:50.771910Z","submitted_at":"2024-01-11T14:09:09Z","title":"SH2: Self-Highlighted Hesitation Helps You Decode More Truthfully","version":4},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2401.05930","snapshot_observed_at":"2026-08-08T22:23:27.974474Z","title":"Sh2: Self-highlighted hesitation helps you decode more truthfully","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":27,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:27.974474Z"},"links":{"cited_paper":"/paper/2401.05930","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:51ba61ad6be64edf0bb3c602bb18775d5601a35d55ed982e822ec5e1f3c51dde","observation_id":"b4a7d798-0b2a-4103-adad-0731163b13f0","resolution":{"observed_at":"2026-08-08T22:23:27.974474Z","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":"2026-08-08T22:23:29.433337Z","title":"Beitrag zur n \\\"a herungsweisen Integration totaler Differentialgleichungen","venue":null,"work_id":"2b3e944d-5b6c-4066-9f92-d94e3646c3f8","year":1901},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":28,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.029424Z"},"links":{"citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:15b2221a5eeeb14987d517d030a233c2c2fcafb46a59bbd9865c1c8225457d41","observation_id":"b918dc1f-940b-4089-a7b8-07212c30b8a9","resolution":{"observed_at":"2026-08-08T22:23:29.436488Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-08T06:32:00.761636+00:00","source":"crossref"},{"observed_at":"2026-08-08T06:31:55.24221+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":"2026-08-08T22:23:29.422479Z","title":"Natural questions: a benchmark for question answering research","venue":null,"work_id":"d674aba2-7c2f-423e-af18-4d5bc08e4238","year":2019},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":29,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.081014Z"},"links":{"citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:8bc130eb0b827ddff84e300f2f2cd140297ca8f84cc65bd85d9370b625997155","observation_id":"7b0c6773-0735-415f-8ee4-8b872093c1e2","resolution":{"observed_at":"2026-08-08T22:23:29.426143Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-08T06:32:00.761636+00:00","source":"crossref"},{"observed_at":"2026-08-08T06:31:55.24221+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-08T22:23:28.171509Z","title":"u ttler, H., Lewis, M., Yih, W.-t., Rockt \\","venue":null,"work_id":null,"year":2020},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":30,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.171509Z"},"links":{"citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:e077bb8fd034507e48ab69c7c7993d88d13780cca5b99169b8a96710479c22f6","observation_id":"6ca2f42b-d331-4589-86bd-2e4ab0f39882","resolution":{"observed_at":"2026-08-08T22:23:28.171509Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2305.11747","last_updated":"2023-10-23T01:49:32Z","snapshot_observed_at":"2026-08-07T22:56:15.588770Z","submitted_at":"2023-05-19T15:36:27Z","title":"HaluEval: A Large-Scale Hallucination Evaluation Benchmark for Large Language Models","version":3},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2305.11747","snapshot_observed_at":"2026-08-08T22:23:28.190876Z","title":"X., Nie, J.-Y., and Wen, J.-R","venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":31,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.190876Z"},"links":{"cited_paper":"/paper/2305.11747","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:cdf9da908569f8252aa326b617cc81e5d1d684bf6a8ab6ce7c051b9d7875c35a","observation_id":"ba465a51-296c-4227-881a-eef3ca3c7eed","resolution":{"observed_at":"2026-08-08T22:23:28.190876Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-08T22:23:28.199961Z","title":"Inference-time intervention: Eliciting truthful answers from a language model","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":32,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.199961Z"},"links":{"citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:60ad3d27b30dc8b8111bcb20d8d7b5c150d8ecbe7924396ae80c55bd3f2eeefa","observation_id":"07e814d5-44aa-43ba-a0b4-04dc2a809c0a","resolution":{"observed_at":"2026-08-08T22:23:28.199961Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2210.15097","last_updated":"2023-07-10T06:08:55Z","snapshot_observed_at":"2026-07-06T14:10:59.567758Z","submitted_at":"2022-10-27T00:58:21Z","title":"Contrastive Decoding: Open-ended Text Generation as Optimization","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2210.15097","snapshot_observed_at":"2026-08-08T22:23:28.203920Z","title":"L., Holtzman, A., Fried, D., Liang, P., Eisner, J., Hashimoto, T., Zettlemoyer, L., and Lewis, M","venue":null,"work_id":null,"year":2022},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":33,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.203920Z"},"links":{"cited_paper":"/paper/2210.15097","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:9ca41399cf0b27c20cb63021e50424a4f001e18c05afd15f1cd3141b008b354a","observation_id":"0b112ed8-4537-4c86-8cbb-ec1d4b7b0808","resolution":{"observed_at":"2026-08-08T22:23:28.203920Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2109.07958","last_updated":"2022-05-08T02:43:02Z","snapshot_observed_at":"2026-08-03T16:26:48.747700Z","submitted_at":"2021-09-08T17:15:27Z","title":"TruthfulQA: Measuring How Models Mimic Human Falsehoods","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2109.07958","snapshot_observed_at":"2026-08-08T22:23:28.209103Z","title":"Truthfulqa: Measuring how models mimic human falsehoods","venue":null,"work_id":null,"year":2021},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":34,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.209103Z"},"links":{"cited_paper":"/paper/2109.07958","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:30b913380973772bb9800f0ab76e4a73e6d1ade7bba4a7a961e7d110ba3f4b62","observation_id":"c2a72bfb-fe41-4d23-b752-3ca5ea9bc3c7","resolution":{"observed_at":"2026-08-08T22:23:28.209103Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2210.02747","last_updated":"2023-02-08T15:46:05Z","snapshot_observed_at":"2026-08-02T18:24:58.914589Z","submitted_at":"2022-10-06T08:32:20Z","title":"Flow Matching for Generative Modeling","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2210.02747","snapshot_observed_at":"2026-08-08T22:23:28.212853Z","title":"T., Ben-Hamu, H., Nickel, M., and Le, M","venue":null,"work_id":null,"year":2022},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":35,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.212853Z"},"links":{"cited_paper":"/paper/2210.02747","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:bd091705dbf51ac074f17095a673da2c877ac363731c7444358ae5e113010076","observation_id":"b0733e87-75d2-4eaf-b163-0ad3baeb2e1e","resolution":{"observed_at":"2026-08-08T22:23:28.212853Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2412.19437","last_updated":"2025-02-18T17:26:38Z","snapshot_observed_at":"2026-07-06T02:11:23.670680Z","submitted_at":"2024-12-27T04:03:16Z","title":"DeepSeek-V3 Technical Report","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2412.19437","snapshot_observed_at":"2026-08-08T22:23:28.265727Z","title":"Deepseek-v3 technical report","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":36,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.265727Z"},"links":{"cited_paper":"/paper/2412.19437","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:1c6397cb5344233651e7d1b807c662b11cde141e3353815180d9f927845bd114","observation_id":"c04a9a4f-0fa6-42ba-ade7-812b9f983853","resolution":{"observed_at":"2026-08-08T22:23:28.265727Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2209.03003","last_updated":"2022-09-07T08:59:55Z","snapshot_observed_at":"2026-07-06T13:49:40.974495Z","submitted_at":"2022-09-07T08:59:55Z","title":"Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2209.03003","snapshot_observed_at":"2026-08-08T22:23:28.371568Z","title":"Flow straight and fast: Learning to generate and transfer data with rectified flow","venue":null,"work_id":null,"year":2022},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":37,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.371568Z"},"links":{"cited_paper":"/paper/2209.03003","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:044c5e1765343344fec6c91741aaa830c38f63213a0279d993d62c7cd9d6b5da","observation_id":"31112765-95d2-4c20-900e-0e8502bb4c9e","resolution":{"observed_at":"2026-08-08T22:23:28.371568Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.01509","last_updated":"2024-06-09T13:07:50Z","snapshot_observed_at":"2026-08-02T03:54:54.655157Z","submitted_at":"2024-03-03T13:14:47Z","title":"Fantastic Semantics and Where to Find Them: Investigating Which Layers of Generative LLMs Reflect Lexical Semantics","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.01509","snapshot_observed_at":"2026-08-08T22:23:28.406670Z","title":"Fantastic semantics and where to find them: Investigating which layers of generative llms reflect lexical semantics","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":38,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.406670Z"},"links":{"cited_paper":"/paper/2403.01509","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:ca8431d6d5c93aa4f843effea2e93d6a6875f36ff7d2b98e2bbb90dd22cd0665","observation_id":"ef38339d-0b30-41fa-b882-91079f074a33","resolution":{"observed_at":"2026-08-08T22:23:28.406670Z","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":"2026-08-08T22:23:29.399370Z","title":"Probing llms for logical reasoning","venue":null,"work_id":"9668e146-a411-4b7e-8f41-a55e192640a3","year":2024},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":39,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.417749Z"},"links":{"citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:be5289d8dfe47bf94799672aaa7707c3ac3d4f4909773e91c9dca84bbb2df50c","observation_id":"d86a6608-8fc0-4e76-a084-e37eab375ce2","resolution":{"observed_at":"2026-08-08T22:23:29.402763Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-08T06:32:00.761636+00:00","source":"crossref"},{"observed_at":"2026-08-08T06:31:55.24221+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2309.09117","last_updated":"2023-09-29T15:11:49Z","snapshot_observed_at":"2026-08-04T19:24:17.623767Z","submitted_at":"2023-09-17T00:29:32Z","title":"Contrastive Decoding Improves Reasoning in Large Language Models","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2309.09117","snapshot_observed_at":"2026-08-08T22:23:28.424314Z","title":"and Lewis, M","venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":40,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.424314Z"},"links":{"cited_paper":"/paper/2309.09117","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:4ea4a641ea8607a0c0503c0a3fe235ba0a12a76721bf1eb8a5b2a940b9bf1a8f","observation_id":"4f19e940-c736-4ddf-a7e1-fcbdc487e549","resolution":{"observed_at":"2026-08-08T22:23:28.424314Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-08T22:23:28.428397Z","title":"Training language models to follow instructions with human feedback","venue":null,"work_id":null,"year":2022},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":41,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.428397Z"},"links":{"citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:09736d0c38186ad1936226dd5482b332a5d860176b993c1450a4329e892e1b17","observation_id":"be610d12-d9aa-4ca7-adf8-ce4398e469e0","resolution":{"observed_at":"2026-08-08T22:23:28.428397Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2307.15343","last_updated":"2023-10-14T17:04:12Z","snapshot_observed_at":"2026-08-07T02:36:54.737484Z","submitted_at":"2023-07-28T06:43:04Z","title":"Med-HALT: Medical Domain Hallucination Test for Large Language Models","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2307.15343","snapshot_observed_at":"2026-08-08T22:23:28.432815Z","title":"K., and Sankarasubbu, M","venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":42,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.432815Z"},"links":{"cited_paper":"/paper/2307.15343","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:072ab1ba3235ee81e0f93b17750021724472499480971ed6d7c42563e4e51c87","observation_id":"87436ef0-890a-4c96-becb-d13c5b1257e8","resolution":{"observed_at":"2026-08-08T22:23:28.432815Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2312.06681","last_updated":"2024-07-05T15:30:45Z","snapshot_observed_at":"2026-08-07T14:28:06.805424Z","submitted_at":"2023-12-09T04:40:46Z","title":"Steering Llama 2 via Contrastive Activation Addition","version":4},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2312.06681","snapshot_observed_at":"2026-08-08T22:23:28.437300Z","title":null,"venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":43,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.437300Z"},"links":{"cited_paper":"/paper/2312.06681","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:c46c7877b0a18572a94a6d064e35340f50d376e724934e8ad27bca3e12fdf2b7","observation_id":"91ac5a9d-a185-43c2-82d1-08e16e47ddbd","resolution":{"observed_at":"2026-08-08T22:23:28.437300Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-08T22:23:28.441246Z","title":"D., Ermon, S., and Finn, C","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":44,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.441246Z"},"links":{"citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:8f53d57a3d6010de64d35bd7dfc52d4da92ebae4ac3e76e71c728c92b63bea99","observation_id":"597c427e-b712-4dba-8951-9fd641b294a8","resolution":{"observed_at":"2026-08-08T22:23:28.441246Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2309.05922","last_updated":"2023-09-12T02:34:06Z","snapshot_observed_at":"2026-07-06T02:11:23.670680Z","submitted_at":"2023-09-12T02:34:06Z","title":"A Survey of Hallucination in Large Foundation Models","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2309.05922","snapshot_observed_at":"2026-08-08T22:23:28.444572Z","title":"A survey of hallucination in large foundation models","venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":45,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.444572Z"},"links":{"cited_paper":"/paper/2309.05922","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:7f3de557e06833a9c42281785efccb8a622da4b4203bd42846101112948c3f32","observation_id":"46460980-2a9e-4f90-a1b3-e2035a6369c7","resolution":{"observed_at":"2026-08-08T22:23:28.444572Z","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":"2026-08-08T22:23:29.371870Z","title":null,"venue":null,"work_id":"f3eb51b1-b18d-4d7f-921c-eaedc7cc2f61","year":2022},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":46,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.447601Z"},"links":{"citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:9e91605ce5ade3a2a3af2b91719c4b946941d3be137d105e7d2f7a5dcc85f257","observation_id":"456845d6-660b-4a3c-930a-b9e16efecb11","resolution":{"observed_at":"2026-08-08T22:23:29.375574Z","resolver_source":"raw_fallback","status":"unresolved"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-08T06:32:00.761636+00:00","source":"crossref"},{"observed_at":"2026-08-08T06:31:55.24221+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-08T22:23:28.451311Z","title":"U-net: Convolutional networks for biomedical image segmentation","venue":null,"work_id":null,"year":2015},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":47,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.451311Z"},"links":{"citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:460f8a8bae71e3046d167b320732818b51f538a06d564ea5d6a34b48fbd24d52","observation_id":"b5ad759b-1650-4294-b34c-c665fd51e3e2","resolution":{"observed_at":"2026-08-08T22:23:28.451311Z","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":"2026-08-08T22:23:29.320047Z","title":"U ber die numerische aufl \\","venue":null,"work_id":"a0eb8d32-218b-4754-8c42-9a4b89dfc89b","year":null},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":48,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.454488Z"},"links":{"citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:d46f33a0b5e2f7c202cc6860906350753d5080e87c9061260d3489226ae532f5","observation_id":"0be1d27a-1b33-46ad-a09b-37d1f57a3e82","resolution":{"observed_at":"2026-08-08T22:23:29.357856Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-08T06:32:00.761636+00:00","source":"crossref"},{"observed_at":"2026-08-08T06:31:55.24221+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2004.04696","last_updated":"2020-05-21T16:53:47Z","snapshot_observed_at":"2026-08-08T18:39:53.307653Z","submitted_at":"2020-04-09T17:26:52Z","title":"BLEURT: Learning Robust Metrics for Text Generation","version":5},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2004.04696","snapshot_observed_at":"2026-08-08T22:23:28.458217Z","title":null,"venue":null,"work_id":null,"year":2004},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":49,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.458217Z"},"links":{"cited_paper":"/paper/2004.04696","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:5d3bd4334aa24fa2a85857aeaeff0ea4c3b3dff322a833b1bcd436b793b06dcc","observation_id":"daac12ba-004d-42f7-9d22-8bc25331a672","resolution":{"observed_at":"2026-08-08T22:23:28.458217Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"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":"2026-08-08T22:23:28.462488Z","title":"S., Love, J., et al","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":50,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.462488Z"},"links":{"cited_paper":"/paper/2403.08295","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:0eb695b17a9e668676db00fc5e55be1c186a8fb2e2fb48e3dbb38b244bf3c0da","observation_id":"a62c728a-fea2-4da8-aee2-f0ee5cc5c77d","resolution":{"observed_at":"2026-08-08T22:23:28.462488Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2311.08401","last_updated":"2023-11-14T18:59:15Z","snapshot_observed_at":"2026-08-04T04:20:13.850943Z","submitted_at":"2023-11-14T18:59:15Z","title":"Fine-tuning Language Models for Factuality","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2311.08401","snapshot_observed_at":"2026-08-08T22:23:28.466620Z","title":"D., and Finn, C","venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":51,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.466620Z"},"links":{"cited_paper":"/paper/2311.08401","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:011e54a79df252043fb5373dc469b946a7641c5012eea7a1fcd6db7404c0cfeb","observation_id":"4fe22985-3b68-4fd6-939a-1d25156a0fd1","resolution":{"observed_at":"2026-08-08T22:23:28.466620Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2307.09288","last_updated":"2023-07-19T17:08:59Z","snapshot_observed_at":"2026-08-07T12:56:43.323460Z","submitted_at":"2023-07-18T14:31:57Z","title":"Llama 2: Open Foundation and Fine-Tuned Chat Models","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2307.09288","snapshot_observed_at":"2026-08-08T22:23:28.470232Z","title":"Llama 2: Open foundation and fine-tuned chat models","venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":52,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.470232Z"},"links":{"cited_paper":"/paper/2307.09288","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:ac3d3fb80939f2f72ce88b53d8d39a661117f3d96d9805ddede96062eb0782fd","observation_id":"84c669bc-1a15-4dce-a263-8e05c6ab312a","resolution":{"observed_at":"2026-08-08T22:23:28.470232Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-08T22:23:28.500981Z","title":"Attention is all you need","venue":null,"work_id":null,"year":2017},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":53,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.500981Z"},"links":{"citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:81d0020f100096019496bd1f4cc151d667d4b25ad9e991314bfc1e6585060217","observation_id":"f93ee7fb-d2fb-444f-b33c-a001ece16d12","resolution":{"observed_at":"2026-08-08T22:23:28.500981Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"1809.09600","last_updated":"2018-09-25T17:28:20Z","snapshot_observed_at":"2026-07-31T11:19:06.421362Z","submitted_at":"2018-09-25T17:28:20Z","title":"HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"1809.09600","snapshot_observed_at":"2026-08-08T22:23:28.556369Z","title":"W., Salakhutdinov, R., and Manning, C","venue":null,"work_id":null,"year":2018},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":54,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.556369Z"},"links":{"cited_paper":"/paper/1809.09600","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:5aa35c6e8c7fc961f13ef2c92672153f71a5db26eff76f038320a798eeb6a6d5","observation_id":"dccf0263-ad27-449d-9892-7fc9035d1282","resolution":{"observed_at":"2026-08-08T22:23:28.556369Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2402.17811","last_updated":"2024-06-05T11:15:04Z","snapshot_observed_at":"2026-07-06T17:36:27.931373Z","submitted_at":"2024-02-27T14:45:04Z","title":"TruthX: Alleviating Hallucinations by Editing Large Language Models in Truthful Space","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2402.17811","snapshot_observed_at":"2026-08-08T22:23:28.595116Z","title":"Truthx: Alleviating hallucinations by editing large language models in truthful space","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":55,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.595116Z"},"links":{"cited_paper":"/paper/2402.17811","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:60a4d66710d475591fd94f2add56334d24d2cdbb7291a90c53031d3e7f56dd57","observation_id":"190d6256-a71b-4efb-a652-648797113d44","resolution":{"observed_at":"2026-08-08T22:23:28.595116Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2312.15710","last_updated":"2024-03-11T07:50:05Z","snapshot_observed_at":"2026-07-06T17:07:52.197869Z","submitted_at":"2023-12-25T12:32:49Z","title":"Alleviating Hallucinations of Large Language Models through Induced Hallucinations","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2312.15710","snapshot_observed_at":"2026-08-08T22:23:28.641619Z","title":"Alleviating hallucinations of large language models through induced hallucinations","venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":56,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.641619Z"},"links":{"cited_paper":"/paper/2312.15710","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:bb5f0f89afee7a9349dde976825b918cc302c516662d543c0d1cfbfee1ac407b","observation_id":"d18f4ee2-97aa-4236-892c-8cd7d6b99b47","resolution":{"observed_at":"2026-08-08T22:23:28.641619Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2310.01405","last_updated":"2025-03-03T06:14:14Z","snapshot_observed_at":"2026-07-06T16:26:38.284922Z","submitted_at":"2023-10-02T17:59:07Z","title":"Representation Engineering: A Top-Down Approach to AI Transparency","version":4},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2310.01405","snapshot_observed_at":"2026-08-08T22:23:28.678697Z","title":"Representation engineering: A top-down approach to ai transparency","venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":57,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.678697Z"},"links":{"cited_paper":"/paper/2310.01405","citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:b9043943232641e9f1dfc0343500309b19f2365e9cedb95ce76735609b6088a1","observation_id":"4e2be219-7c2c-46d3-906e-f41177bee29e","resolution":{"observed_at":"2026-08-08T22:23:28.678697Z","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":"2026-08-08T22:23:29.268641Z","title":"Z., Fredrikson, M., and Hendrycks, D","venue":null,"work_id":"eac6c31e-adfa-472e-a9db-bf6af1d7a48c","year":2024},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":58,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.721264Z"},"links":{"citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:0b722f65acba4a8faf2ad591a62bf35c9ba5beb8bc49d2861b9defe913fe9814","observation_id":"7490333d-c274-4b8c-b62c-e5a8ccffff81","resolution":{"observed_at":"2026-08-08T22:23:29.295766Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-08T06:32:00.761636+00:00","source":"crossref"},{"observed_at":"2026-08-08T06:31:55.24221+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-08T22:23:28.738176Z","title":"write newline","venue":null,"work_id":null,"year":null},"citing_paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","version":1},"reference_index":59,"source":"arxiv_source","source_observed_at":"2026-08-08T22:23:28.738176Z"},"links":{"citing_paper":"/paper/2502.04556"},"observation_digest":"sha256:265d99e8bc94f570b7f33cee9e8a23441a90358ce616a6853d3a5d6c20c229de","observation_id":"83baf9cc-db12-4b9f-97b5-a837c9ef14fb","resolution":{"observed_at":"2026-08-08T22:23:28.738176Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}}],"paper":{"arxiv_id":"2502.04556","last_updated":"2025-02-06T23:10:14Z","latest_version":1,"primary_category":"cs.CL","snapshot_observed_at":"2026-08-08T22:32:29.549373Z","submitted_at":"2025-02-06T23:10:14Z","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction"},"reference_resolution":{"displayed":59,"state_counts":{"malformed_identifier":0,"metadata_mismatch":0,"parse_uncertain":0,"unresolved":50,"verified_exact":1,"verified_fuzzy":8},"total_outbound_references":59},"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":"2026-08-08T06:32:00.761636+00:00","source":"crossref"},{"observed_at":"2026-08-08T06:31:55.24221+00:00","source":"retraction_watch"}],"thesis":"As of 8 August 2026, this Paper Citation Record lists 59 of 59 outbound references and 4 inbound Pith citation observations for arXiv:2502.04556."}