{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:C7R5UVHJBG6DS2TXAG7XXFCWIO","short_pith_number":"pith:C7R5UVHJ","schema_version":"1.0","canonical_sha256":"17e3da54e909bc396a7701bf7b945643ac030298cc0c4b41d5cfbb0aaebc74dc","source":{"kind":"arxiv","id":"2508.13152","version":1},"attestation_state":"computed","paper":{"title":"RepreGuard: Detecting LLM-Generated Text by Revealing Hidden Representation Patterns","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Derek F. Wong, Di Wang, Junchao Wu, Lidia S. Chao, Min Yang, Runzhe Zhan, Shu Yang, Xin Chen, Zeyu Wu, Ziyang Luo","submitted_at":"2025-08-18T17:59:15Z","abstract_excerpt":"Detecting content generated by large language models (LLMs) is crucial for preventing misuse and building trustworthy AI systems. Although existing detection methods perform well, their robustness in out-of-distribution (OOD) scenarios is still lacking. In this paper, we hypothesize that, compared to features used by existing detection methods, the internal representations of LLMs contain more comprehensive and raw features that can more effectively capture and distinguish the statistical pattern differences between LLM-generated texts (LGT) and human-written texts (HWT). We validated this hyp"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2508.13152","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-08-18T17:59:15Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"179b888fe9614a39a992ba7165445e2425e9a60caf7286d55482507cdf1bba13","abstract_canon_sha256":"21afe1853d6c56184e49e5c05024c99081d79227469c1425c51b81499733fa90"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:55:33.297448Z","signature_b64":"UVSnJylBKaptkRMcqWZqLETAFhsSRJ7UznaBvZn2Y3bQyhWWvVfP8yoOcDV5Pp2PZT0PewK6H7ktSDvxSrDtDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"17e3da54e909bc396a7701bf7b945643ac030298cc0c4b41d5cfbb0aaebc74dc","last_reissued_at":"2026-07-05T11:55:33.296962Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:55:33.296962Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"RepreGuard: Detecting LLM-Generated Text by Revealing Hidden Representation Patterns","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Derek F. Wong, Di Wang, Junchao Wu, Lidia S. Chao, Min Yang, Runzhe Zhan, Shu Yang, Xin Chen, Zeyu Wu, Ziyang Luo","submitted_at":"2025-08-18T17:59:15Z","abstract_excerpt":"Detecting content generated by large language models (LLMs) is crucial for preventing misuse and building trustworthy AI systems. Although existing detection methods perform well, their robustness in out-of-distribution (OOD) scenarios is still lacking. In this paper, we hypothesize that, compared to features used by existing detection methods, the internal representations of LLMs contain more comprehensive and raw features that can more effectively capture and distinguish the statistical pattern differences between LLM-generated texts (LGT) and human-written texts (HWT). We validated this hyp"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.13152","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2508.13152/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2508.13152","created_at":"2026-07-05T11:55:33.297020+00:00"},{"alias_kind":"arxiv_version","alias_value":"2508.13152v1","created_at":"2026-07-05T11:55:33.297020+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.13152","created_at":"2026-07-05T11:55:33.297020+00:00"},{"alias_kind":"pith_short_12","alias_value":"C7R5UVHJBG6D","created_at":"2026-07-05T11:55:33.297020+00:00"},{"alias_kind":"pith_short_16","alias_value":"C7R5UVHJBG6DS2TX","created_at":"2026-07-05T11:55:33.297020+00:00"},{"alias_kind":"pith_short_8","alias_value":"C7R5UVHJ","created_at":"2026-07-05T11:55:33.297020+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.06315","citing_title":"LLM Self-Recognition: Steering and Retrieving Activation Signatures","ref_index":70,"is_internal_anchor":false},{"citing_arxiv_id":"2605.22654","citing_title":"Seeing the Poem: Image-Semantic Detection of AI-Generated Modern Chinese Poetry with MLLMs","ref_index":84,"is_internal_anchor":false},{"citing_arxiv_id":"2605.16107","citing_title":"Multi-Level Contextual Token Relation Modeling for Machine-Generated Text Detection","ref_index":39,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/C7R5UVHJBG6DS2TXAG7XXFCWIO","json":"https://pith.science/pith/C7R5UVHJBG6DS2TXAG7XXFCWIO.json","graph_json":"https://pith.science/api/pith-number/C7R5UVHJBG6DS2TXAG7XXFCWIO/graph.json","events_json":"https://pith.science/api/pith-number/C7R5UVHJBG6DS2TXAG7XXFCWIO/events.json","paper":"https://pith.science/paper/C7R5UVHJ"},"agent_actions":{"view_html":"https://pith.science/pith/C7R5UVHJBG6DS2TXAG7XXFCWIO","download_json":"https://pith.science/pith/C7R5UVHJBG6DS2TXAG7XXFCWIO.json","view_paper":"https://pith.science/paper/C7R5UVHJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2508.13152&json=true","fetch_graph":"https://pith.science/api/pith-number/C7R5UVHJBG6DS2TXAG7XXFCWIO/graph.json","fetch_events":"https://pith.science/api/pith-number/C7R5UVHJBG6DS2TXAG7XXFCWIO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/C7R5UVHJBG6DS2TXAG7XXFCWIO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/C7R5UVHJBG6DS2TXAG7XXFCWIO/action/storage_attestation","attest_author":"https://pith.science/pith/C7R5UVHJBG6DS2TXAG7XXFCWIO/action/author_attestation","sign_citation":"https://pith.science/pith/C7R5UVHJBG6DS2TXAG7XXFCWIO/action/citation_signature","submit_replication":"https://pith.science/pith/C7R5UVHJBG6DS2TXAG7XXFCWIO/action/replication_record"}},"created_at":"2026-07-05T11:55:33.297020+00:00","updated_at":"2026-07-05T11:55:33.297020+00:00"}