{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:5N3QCE3ZQBCYWGGJPJRSCDXONX","short_pith_number":"pith:5N3QCE3Z","schema_version":"1.0","canonical_sha256":"eb7701137980458b18c97a63210eee6dc5e7baf36ae06fb2dad906d2211159ea","source":{"kind":"arxiv","id":"2503.17965","version":1},"attestation_state":"computed","paper":{"title":"Understanding the Effects of RLHF on the Quality and Detectability of LLM-Generated Texts","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Arkaitz Zubiaga, Beining Xu","submitted_at":"2025-03-23T07:03:10Z","abstract_excerpt":"Large Language Models (LLMs) have demonstrated exceptional performance on a range of downstream NLP tasks by generating text that closely resembles human writing. However, the ease of achieving this similarity raises concerns from potential malicious uses at scale by bad actors, as LLM-generated text becomes increasingly difficult to discern from human text. Although detection methods have been developed to address this issue, bad actors can further manipulate LLM-generated texts to make them less detectable. In this work, we study how further editing texts with Reinforcement Learning from Hum"},"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":"2503.17965","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-03-23T07:03:10Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"7374ebecac16c47b1b1e15dc887c578e49ea585301ef09dcbfb203b442f9cea1","abstract_canon_sha256":"44a78655cf27f654f211e21bc02e024643e0ecd211f57b465333ffb45ac10d3c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:37:59.536879Z","signature_b64":"uGbYmiZOgBoW1xtkeJO3K7zI/Fz+mct3QuPwbxtIoPlYm1R3QTtgyHspdlTMyD5DeUyD2Q8wrTZ9cVkTh+wuCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"eb7701137980458b18c97a63210eee6dc5e7baf36ae06fb2dad906d2211159ea","last_reissued_at":"2026-07-05T10:37:59.536303Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:37:59.536303Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Understanding the Effects of RLHF on the Quality and Detectability of LLM-Generated Texts","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Arkaitz Zubiaga, Beining Xu","submitted_at":"2025-03-23T07:03:10Z","abstract_excerpt":"Large Language Models (LLMs) have demonstrated exceptional performance on a range of downstream NLP tasks by generating text that closely resembles human writing. However, the ease of achieving this similarity raises concerns from potential malicious uses at scale by bad actors, as LLM-generated text becomes increasingly difficult to discern from human text. Although detection methods have been developed to address this issue, bad actors can further manipulate LLM-generated texts to make them less detectable. In this work, we study how further editing texts with Reinforcement Learning from Hum"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.17965","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/2503.17965/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":"2503.17965","created_at":"2026-07-05T10:37:59.536372+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.17965v1","created_at":"2026-07-05T10:37:59.536372+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.17965","created_at":"2026-07-05T10:37:59.536372+00:00"},{"alias_kind":"pith_short_12","alias_value":"5N3QCE3ZQBCY","created_at":"2026-07-05T10:37:59.536372+00:00"},{"alias_kind":"pith_short_16","alias_value":"5N3QCE3ZQBCYWGGJ","created_at":"2026-07-05T10:37:59.536372+00:00"},{"alias_kind":"pith_short_8","alias_value":"5N3QCE3Z","created_at":"2026-07-05T10:37:59.536372+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5N3QCE3ZQBCYWGGJPJRSCDXONX","json":"https://pith.science/pith/5N3QCE3ZQBCYWGGJPJRSCDXONX.json","graph_json":"https://pith.science/api/pith-number/5N3QCE3ZQBCYWGGJPJRSCDXONX/graph.json","events_json":"https://pith.science/api/pith-number/5N3QCE3ZQBCYWGGJPJRSCDXONX/events.json","paper":"https://pith.science/paper/5N3QCE3Z"},"agent_actions":{"view_html":"https://pith.science/pith/5N3QCE3ZQBCYWGGJPJRSCDXONX","download_json":"https://pith.science/pith/5N3QCE3ZQBCYWGGJPJRSCDXONX.json","view_paper":"https://pith.science/paper/5N3QCE3Z","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.17965&json=true","fetch_graph":"https://pith.science/api/pith-number/5N3QCE3ZQBCYWGGJPJRSCDXONX/graph.json","fetch_events":"https://pith.science/api/pith-number/5N3QCE3ZQBCYWGGJPJRSCDXONX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5N3QCE3ZQBCYWGGJPJRSCDXONX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5N3QCE3ZQBCYWGGJPJRSCDXONX/action/storage_attestation","attest_author":"https://pith.science/pith/5N3QCE3ZQBCYWGGJPJRSCDXONX/action/author_attestation","sign_citation":"https://pith.science/pith/5N3QCE3ZQBCYWGGJPJRSCDXONX/action/citation_signature","submit_replication":"https://pith.science/pith/5N3QCE3ZQBCYWGGJPJRSCDXONX/action/replication_record"}},"created_at":"2026-07-05T10:37:59.536372+00:00","updated_at":"2026-07-05T10:37:59.536372+00:00"}