{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:S7OX6BCZHQ5VEPPWB2XFZLXYVQ","short_pith_number":"pith:S7OX6BCZ","schema_version":"1.0","canonical_sha256":"97dd7f04593c3b523df60eae5caef8ac28b2c043d9c639e9aa880886a43aef62","source":{"kind":"arxiv","id":"2305.14757","version":2},"attestation_state":"computed","paper":{"title":"Psychological Metrics for Dialog System Evaluation","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Farhan Ahmed, Gary Pan, H. Andrew Schwartz, Joao Sedoc, Lyle H. Ungar, Salvatore Giorgi, Shalaka Vaidya, Shreya Havaldar, Zuhaib Akhtar","submitted_at":"2023-05-24T06:02:32Z","abstract_excerpt":"We present metrics for evaluating dialog systems through a psychologically-grounded \"human\" lens in which conversational agents express a diversity of both states (e.g., emotion) and traits (e.g., personality), just as people do. We present five interpretable metrics from established psychology that are fundamental to human communication and relationships: emotional entropy, linguistic style and emotion matching, agreeableness, and empathy. These metrics can be applied (1) across dialogs and (2) on turns within dialogs. The psychological metrics are compared against seven state-of-the-art trad"},"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":"2305.14757","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2023-05-24T06:02:32Z","cross_cats_sorted":[],"title_canon_sha256":"816d56cd89db2a4c57cd60df04ccbf416113c0e99d2b4e28df5f80c356de1d8e","abstract_canon_sha256":"5b9e44925fef16e19c1e481a308bfa0b6d52d64e561166489e8074ef619156c4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:51:10.139786Z","signature_b64":"l6R5la5ZMmgzzrthfzKzt1BveCKDWrjsDH8PuDmnoz+axvaq/o8NSYzWVS29tiJ++Jx3WbuP2du3cKJG0kbnDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"97dd7f04593c3b523df60eae5caef8ac28b2c043d9c639e9aa880886a43aef62","last_reissued_at":"2026-07-05T06:51:10.139341Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:51:10.139341Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Psychological Metrics for Dialog System Evaluation","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Farhan Ahmed, Gary Pan, H. Andrew Schwartz, Joao Sedoc, Lyle H. Ungar, Salvatore Giorgi, Shalaka Vaidya, Shreya Havaldar, Zuhaib Akhtar","submitted_at":"2023-05-24T06:02:32Z","abstract_excerpt":"We present metrics for evaluating dialog systems through a psychologically-grounded \"human\" lens in which conversational agents express a diversity of both states (e.g., emotion) and traits (e.g., personality), just as people do. We present five interpretable metrics from established psychology that are fundamental to human communication and relationships: emotional entropy, linguistic style and emotion matching, agreeableness, and empathy. These metrics can be applied (1) across dialogs and (2) on turns within dialogs. The psychological metrics are compared against seven state-of-the-art trad"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.14757","kind":"arxiv","version":2},"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/2305.14757/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":"2305.14757","created_at":"2026-07-05T06:51:10.139407+00:00"},{"alias_kind":"arxiv_version","alias_value":"2305.14757v2","created_at":"2026-07-05T06:51:10.139407+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.14757","created_at":"2026-07-05T06:51:10.139407+00:00"},{"alias_kind":"pith_short_12","alias_value":"S7OX6BCZHQ5V","created_at":"2026-07-05T06:51:10.139407+00:00"},{"alias_kind":"pith_short_16","alias_value":"S7OX6BCZHQ5VEPPW","created_at":"2026-07-05T06:51:10.139407+00:00"},{"alias_kind":"pith_short_8","alias_value":"S7OX6BCZ","created_at":"2026-07-05T06:51:10.139407+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/S7OX6BCZHQ5VEPPWB2XFZLXYVQ","json":"https://pith.science/pith/S7OX6BCZHQ5VEPPWB2XFZLXYVQ.json","graph_json":"https://pith.science/api/pith-number/S7OX6BCZHQ5VEPPWB2XFZLXYVQ/graph.json","events_json":"https://pith.science/api/pith-number/S7OX6BCZHQ5VEPPWB2XFZLXYVQ/events.json","paper":"https://pith.science/paper/S7OX6BCZ"},"agent_actions":{"view_html":"https://pith.science/pith/S7OX6BCZHQ5VEPPWB2XFZLXYVQ","download_json":"https://pith.science/pith/S7OX6BCZHQ5VEPPWB2XFZLXYVQ.json","view_paper":"https://pith.science/paper/S7OX6BCZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2305.14757&json=true","fetch_graph":"https://pith.science/api/pith-number/S7OX6BCZHQ5VEPPWB2XFZLXYVQ/graph.json","fetch_events":"https://pith.science/api/pith-number/S7OX6BCZHQ5VEPPWB2XFZLXYVQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/S7OX6BCZHQ5VEPPWB2XFZLXYVQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/S7OX6BCZHQ5VEPPWB2XFZLXYVQ/action/storage_attestation","attest_author":"https://pith.science/pith/S7OX6BCZHQ5VEPPWB2XFZLXYVQ/action/author_attestation","sign_citation":"https://pith.science/pith/S7OX6BCZHQ5VEPPWB2XFZLXYVQ/action/citation_signature","submit_replication":"https://pith.science/pith/S7OX6BCZHQ5VEPPWB2XFZLXYVQ/action/replication_record"}},"created_at":"2026-07-05T06:51:10.139407+00:00","updated_at":"2026-07-05T06:51:10.139407+00:00"}