{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:PBARCWK3REI6OBQQP4PIDPONC3","short_pith_number":"pith:PBARCWK3","schema_version":"1.0","canonical_sha256":"784111595b8911e706107f1e81bdcd16cb47bedf87c1b7faef6bb0994d43b9d8","source":{"kind":"arxiv","id":"2607.27816","version":1},"attestation_state":"computed","paper":{"title":"Beyond Borrowed Histories: Person-Aligned User Simulation for Interactive Role-Playing Evaluation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"2), (2) MetaStone Technology, Beijing, Benfeng Xu (1, China), Hongtao Xie (1) ((1) University of Science, Jie Gao (2), Lingyun Yu (1), Mingxuan Du (1), Technology of China, Yuhang Zhu (1)","submitted_at":"2026-07-30T07:55:35Z","abstract_excerpt":"Role-playing agents (RPAs) have become one of the most important consumer applications of large language models. Users engage in multi-turn conversations with RPAs for experiences such as emotional comfort, making reliable evaluation essential for measuring capability, comparing systems, and guiding further improvement. Existing benchmarks, however, typically require an RPA to continue a fixed dialogue history and then evaluate the continuation using a fixed rubric detached from the user. We identify and empirically demonstrate two limitations of this design. First, an RPA's output is shaped b"},"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":"2607.27816","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-07-30T07:55:35Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"6030ea9aaf2ac46d30dd9c7ddb6db04a3ac9d40c7c5cdf058a5771d0163d869c","abstract_canon_sha256":"91475f8e42228ad9e6cc3bdcf47b3d71b9c0236fb6e1399d0fa45c5f4e8aac47"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"784111595b8911e706107f1e81bdcd16cb47bedf87c1b7faef6bb0994d43b9d8","last_reissued_at":"2026-07-31T01:34:16.285789Z","signature_status":"unsigned_v0","first_computed_at":"2026-07-31T01:34:16.285789Z"},"graph_snapshot":{"paper":{"title":"Beyond Borrowed Histories: Person-Aligned User Simulation for Interactive Role-Playing Evaluation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"2), (2) MetaStone Technology, Beijing, Benfeng Xu (1, China), Hongtao Xie (1) ((1) University of Science, Jie Gao (2), Lingyun Yu (1), Mingxuan Du (1), Technology of China, Yuhang Zhu (1)","submitted_at":"2026-07-30T07:55:35Z","abstract_excerpt":"Role-playing agents (RPAs) have become one of the most important consumer applications of large language models. Users engage in multi-turn conversations with RPAs for experiences such as emotional comfort, making reliable evaluation essential for measuring capability, comparing systems, and guiding further improvement. Existing benchmarks, however, typically require an RPA to continue a fixed dialogue history and then evaluate the continuation using a fixed rubric detached from the user. We identify and empirically demonstrate two limitations of this design. First, an RPA's output is shaped b"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.27816","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/2607.27816/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":"2607.27816","created_at":"2026-07-31T01:34:16.289003+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.27816v1","created_at":"2026-07-31T01:34:16.289003+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.27816","created_at":"2026-07-31T01:34:16.289003+00:00"},{"alias_kind":"pith_short_12","alias_value":"PBARCWK3REI6","created_at":"2026-07-31T01:34:16.289003+00:00"},{"alias_kind":"pith_short_16","alias_value":"PBARCWK3REI6OBQQ","created_at":"2026-07-31T01:34:16.289003+00:00"},{"alias_kind":"pith_short_8","alias_value":"PBARCWK3","created_at":"2026-07-31T01:34:16.289003+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/PBARCWK3REI6OBQQP4PIDPONC3","json":"https://pith.science/pith/PBARCWK3REI6OBQQP4PIDPONC3.json","graph_json":"https://pith.science/api/pith-number/PBARCWK3REI6OBQQP4PIDPONC3/graph.json","events_json":"https://pith.science/api/pith-number/PBARCWK3REI6OBQQP4PIDPONC3/events.json","paper":"https://pith.science/paper/PBARCWK3"},"agent_actions":{"view_html":"https://pith.science/pith/PBARCWK3REI6OBQQP4PIDPONC3","download_json":"https://pith.science/pith/PBARCWK3REI6OBQQP4PIDPONC3.json","view_paper":"https://pith.science/paper/PBARCWK3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.27816&json=true","fetch_graph":"https://pith.science/api/pith-number/PBARCWK3REI6OBQQP4PIDPONC3/graph.json","fetch_events":"https://pith.science/api/pith-number/PBARCWK3REI6OBQQP4PIDPONC3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PBARCWK3REI6OBQQP4PIDPONC3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PBARCWK3REI6OBQQP4PIDPONC3/action/storage_attestation","attest_author":"https://pith.science/pith/PBARCWK3REI6OBQQP4PIDPONC3/action/author_attestation","sign_citation":"https://pith.science/pith/PBARCWK3REI6OBQQP4PIDPONC3/action/citation_signature","submit_replication":"https://pith.science/pith/PBARCWK3REI6OBQQP4PIDPONC3/action/replication_record"}},"created_at":"2026-07-31T01:34:16.289003+00:00","updated_at":"2026-07-31T01:34:16.289003+00:00"}