{"id":"547090f27774efdc6368996f1ecdd9d5466a4904fde83d8da008169dc13a40b7","title":"Fangyuan Liu and coauthors present PsyJev for psychological analysis","argument":"In a paper submitted to arXiv, Fangyuan Liu and coauthors introduce PsyJev, a model for answering questions about psychological information in online self-disclosures. They organize PsyJudgeBench, a benchmark containing varied datasets and questions, and train models to produce answer distributions conditioned on each question. The authors report that their Qwen3.5-based models achieve average test F1 scores of 64.33, 66.21, and 67.56 for the 2B, 4B, and 9B models; they say the 9B model exceeds the strongest external comparison by 2.54 points. Escape Meridian's reading: these results describe performance on the benchmark, not established reliability in real-world psychological or clinical use. How well does PsyJev perform on self-disclosures outside the benchmark, especially when errors could affect people’s care or assessment?","topicIds":["intelligence-and-discovery"],"domain":"science","claimType":"attributed reading","limits":"The evidence is the arXiv abstract; it does not provide the underlying evaluation details or establish how the models perform in real-world psychological or clinical settings.","speaker":{"name":"Liu, Fangyuan, Zhao, Sirui, Wang, Xiaohao, Cheng, Guo, Zhang, Peng, Chen, Enhong","roleAtSource":"Article author"},"source":{"title":"PsyJev: Psychological Analysis via System-1 Judgment","publisher":"arXiv cs.AI","format":"Article","url":"https://arxiv.org/abs/2610.36689","publishedAt":"2026-10-27","dateBasis":"Date basis not provided."},"evidence":[{"text":"The authors introduce PsyJudgeBench, a benchmark with 13 annotated datasets and 82 questions across six psychological areas.","excerpt":"PsyJudgeBench, comprising 13 annotated datasets and 82 questions across six psychological areas","url":"https://arxiv.org/abs/2610.36689","relationship":"Source-stated"},{"text":"The authors introduce PsyJev, which produces answer distributions specified by each request.","excerpt":"We introduce PsyJev, a System-1 model that directly produces distributions over answers specified in each request.","url":"https://arxiv.org/abs/2610.36689","relationship":"Source-stated"},{"text":"The authors report average test F1 scores of 64.33, 66.21, and 67.56 for their Qwen3.5-based 2B, 4B, and 9B models.","excerpt":"The Qwen3.5-based 2B, 4B, and 9B models achieve 64.33, 66.21, and 67.56 average test F1, respectively.","url":"https://arxiv.org/abs/2610.36689","relationship":"Source-stated"},{"text":"The authors report that the 9B model exceeds the strongest external comparison by 2.54 points.","excerpt":"The 9B model exceeds the strongest external comparison by 2.54 points.","url":"https://arxiv.org/abs/2610.36689","relationship":"Source-stated"},{"text":"The abstract reports benchmark test results but does not establish real-world psychological or clinical reliability.","excerpt":"Further analyses examine question-conditioned judgment, inference cost, and probability quality, complemented by qualitative comparisons across models.","url":"https://arxiv.org/abs/2610.36689","relationship":"Escape Meridian interpretation"}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2610.36689","checkedAt":"2026-11-06T15:32:10.001Z","items":[],"note":"No suitable media was found by the supported page extraction."},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"How well does PsyJev perform on self-disclosures outside the benchmark, especially when errors could affect people’s care or assessment?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"Fangyuan Liu and coauthors present PsyJev for psychological analysis","summary":"In a paper submitted to arXiv, Fangyuan Liu and coauthors introduce PsyJev, a model for answering questions about psychological information in online self-disclosures. They organize PsyJudgeBench, a benchmark containing varied datasets and questions, and train models to produce answer distributions conditioned on each question. The authors report that their Qwen3.5-based models achieve average test F1 scores of 64.33, 66.21, and 67.56 for the 2B, 4B, and 9B models; they say the 9B model exceeds the strongest external comparison by 2.54 points. Escape Meridian's reading: these results describe performance on the benchmark, not established reliability in real-world psychological or clinical use. How well does PsyJev perform on self-disclosures outside the benchmark, especially when errors could affect people’s care or assessment?","question":"How well does PsyJev perform on self-disclosures outside the benchmark, especially when errors could affect people’s care or assessment?","claims":[{"text":"The authors introduce PsyJudgeBench, a benchmark with 13 annotated datasets and 82 questions across six psychological areas.","excerpt":"PsyJudgeBench, comprising 13 annotated datasets and 82 questions across six psychological areas","support":"source-stated"},{"text":"The authors introduce PsyJev, which produces answer distributions specified by each request.","excerpt":"We introduce PsyJev, a System-1 model that directly produces distributions over answers specified in each request.","support":"source-stated"},{"text":"The authors report average test F1 scores of 64.33, 66.21, and 67.56 for their Qwen3.5-based 2B, 4B, and 9B models.","excerpt":"The Qwen3.5-based 2B, 4B, and 9B models achieve 64.33, 66.21, and 67.56 average test F1, respectively.","support":"source-stated"},{"text":"The authors report that the 9B model exceeds the strongest external comparison by 2.54 points.","excerpt":"The 9B model exceeds the strongest external comparison by 2.54 points.","support":"source-stated"},{"text":"The abstract reports benchmark test results but does not establish real-world psychological or clinical reliability.","excerpt":"Further analyses examine question-conditioned judgment, inference cost, and probability quality, complemented by qualitative comparisons across models.","support":"interpretation"}],"limitations":["The evidence is the arXiv abstract; it does not provide the underlying evaluation details or establish how the models perform in real-world psychological or clinical settings."],"mediaId":"402625908a648d0345f9eff1194dde1bdfc76ab07adeeff2da6c04edf2211d25","correctionReason":null,"id":"547090f27774efdc6368996f1ecdd9d5466a4904fde83d8da008169dc13a40b7","publicationId":"5e92e232-898b-4af0-8d35-06e9a9733366","version":1,"sourceUrl":"https://arxiv.org/abs/2610.36689","sourceTitle":"PsyJev: Psychological Analysis via System-1 Judgment","publisher":"arXiv cs.AI","publishedAt":null,"originalPublished":{"value":"2026-10-27","precision":"day","basis":"publisher_citation_date"},"firstObservedAt":"2026-10-28T05:28:35.145Z","createdAt":"2026-11-06T11:16:56.477Z","historical":true,"entityIds":["arxiv"],"topicIds":["intelligence-and-discovery"],"media":[{"id":"402625908a648d0345f9eff1194dde1bdfc76ab07adeeff2da6c04edf2211d25","kind":"article","url":"https://arxiv.org/abs/2610.36689","sourceUrl":"https://arxiv.org/abs/2610.36689","label":"Read original","provenance":"source_link","verification":{"status":"verified","checkedAt":"2026-10-30T00:32:18.009Z","method":"acquired_source","contentType":"text/html"},"recommendationReason":"Recommended by Jev for understanding this source; checked by Luna."}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2610.36689","checkedAt":"2026-11-06T15:32:10.001Z","items":[],"note":"No suitable media was found by the supported page extraction."},"editorialDecision":{"outcome":"accept","reasonCodes":["material_development"],"rubricVersion":"stream-admission-v5","mediaId":"402625908a648d0345f9eff1194dde1bdfc76ab07adeeff2da6c04edf2211d25","confidence":0.84,"provider":"typesafe","model":"jev-1.13.0","inputHash":"ef2063d93a0568f6c0c469c52235411296758cf07b27e2235a893f764f6adbce","decidedAt":"2026-10-30T00:32:21.545Z"},"provenance":{"textHash":"5ee63d29a5ba85dfe2b49b811e96cff57eb38f1d55120a88a1727cfd9fb514ca","observationId":"d46f6e99b512644709ddb7adde0ebe971feae8ce145b1fd43748a52a286881b0","retrievedAt":"2026-10-30T00:32:17.477Z","kind":"publisher-page","preparer":{"model":"gpt-6-luna","runId":"f660a074-1687-4610-8955-b741784f4dd6"},"reviewer":{"model":"gpt-6-luna","runId":"3a453453-37d5-4b01-8975-55239cf51dc7"}},"audit":{"status":"uncertain","reviewedAt":"2026-11-06T11:16:56.477Z","model":"anthropic/claude-opus-5.5"},"sourceAttribution":{"authors":[{"name":"Liu, Fangyuan","url":null,"kind":"unknown"},{"name":"Zhao, Sirui","url":null,"kind":"unknown"},{"name":"Wang, Xiaohao","url":null,"kind":"unknown"},{"name":"Cheng, Guo","url":null,"kind":"unknown"},{"name":"Zhang, Peng","url":null,"kind":"unknown"},{"name":"Chen, Enhong","url":null,"kind":"unknown"}],"authorRole":"article-author","hosts":[],"status":"stated","authorRoster":{"truncated":false,"omitted":0,"assertionGroupsOmitted":0},"participation":[{"actor":{"name":"Liu, Fangyuan","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Liu, Fangyuan","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2610.36689","locator":"head/meta[name=author|citation_author]"},"assertionId":"91394d56ae017d2dfb950aa90114624428e1b727781dfba7234195b5a56f5bf5:participation:0","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-10-30T00:32:17.477Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"91394d56ae017d2dfb950aa90114624428e1b727781dfba7234195b5a56f5bf5","sourceUrl":"https://arxiv.org/abs/2610.36689","contentSha256":"06ad2c02a0e0a0b7ca9f9c3948d8210cd96fc1cb67c0c405bfa8d20f765c0f79","textHash":"5ee63d29a5ba85dfe2b49b811e96cff57eb38f1d55120a88a1727cfd9fb514ca","observedAt":"2026-10-30T00:32:17.477Z"}},{"actor":{"name":"Zhao, Sirui","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Zhao, Sirui","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2610.36689","locator":"head/meta[name=author|citation_author]"},"assertionId":"91394d56ae017d2dfb950aa90114624428e1b727781dfba7234195b5a56f5bf5:participation:1","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-10-30T00:32:17.477Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"91394d56ae017d2dfb950aa90114624428e1b727781dfba7234195b5a56f5bf5","sourceUrl":"https://arxiv.org/abs/2610.36689","contentSha256":"06ad2c02a0e0a0b7ca9f9c3948d8210cd96fc1cb67c0c405bfa8d20f765c0f79","textHash":"5ee63d29a5ba85dfe2b49b811e96cff57eb38f1d55120a88a1727cfd9fb514ca","observedAt":"2026-10-30T00:32:17.477Z"}},{"actor":{"name":"Wang, Xiaohao","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Wang, Xiaohao","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2610.36689","locator":"head/meta[name=author|citation_author]"},"assertionId":"91394d56ae017d2dfb950aa90114624428e1b727781dfba7234195b5a56f5bf5:participation:2","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-10-30T00:32:17.477Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"91394d56ae017d2dfb950aa90114624428e1b727781dfba7234195b5a56f5bf5","sourceUrl":"https://arxiv.org/abs/2610.36689","contentSha256":"06ad2c02a0e0a0b7ca9f9c3948d8210cd96fc1cb67c0c405bfa8d20f765c0f79","textHash":"5ee63d29a5ba85dfe2b49b811e96cff57eb38f1d55120a88a1727cfd9fb514ca","observedAt":"2026-10-30T00:32:17.477Z"}},{"actor":{"name":"Cheng, Guo","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Cheng, Guo","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2610.36689","locator":"head/meta[name=author|citation_author]"},"assertionId":"91394d56ae017d2dfb950aa90114624428e1b727781dfba7234195b5a56f5bf5:participation:3","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-10-30T00:32:17.477Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"91394d56ae017d2dfb950aa90114624428e1b727781dfba7234195b5a56f5bf5","sourceUrl":"https://arxiv.org/abs/2610.36689","contentSha256":"06ad2c02a0e0a0b7ca9f9c3948d8210cd96fc1cb67c0c405bfa8d20f765c0f79","textHash":"5ee63d29a5ba85dfe2b49b811e96cff57eb38f1d55120a88a1727cfd9fb514ca","observedAt":"2026-10-30T00:32:17.477Z"}},{"actor":{"name":"Zhang, Peng","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Zhang, Peng","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2610.36689","locator":"head/meta[name=author|citation_author]"},"assertionId":"91394d56ae017d2dfb950aa90114624428e1b727781dfba7234195b5a56f5bf5:participation:4","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-10-30T00:32:17.477Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"91394d56ae017d2dfb950aa90114624428e1b727781dfba7234195b5a56f5bf5","sourceUrl":"https://arxiv.org/abs/2610.36689","contentSha256":"06ad2c02a0e0a0b7ca9f9c3948d8210cd96fc1cb67c0c405bfa8d20f765c0f79","textHash":"5ee63d29a5ba85dfe2b49b811e96cff57eb38f1d55120a88a1727cfd9fb514ca","observedAt":"2026-10-30T00:32:17.477Z"}},{"actor":{"name":"Chen, Enhong","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Chen, Enhong","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2610.36689","locator":"head/meta[name=author|citation_author]"},"assertionId":"91394d56ae017d2dfb950aa90114624428e1b727781dfba7234195b5a56f5bf5:participation:5","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-10-30T00:32:17.477Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"91394d56ae017d2dfb950aa90114624428e1b727781dfba7234195b5a56f5bf5","sourceUrl":"https://arxiv.org/abs/2610.36689","contentSha256":"06ad2c02a0e0a0b7ca9f9c3948d8210cd96fc1cb67c0c405bfa8d20f765c0f79","textHash":"5ee63d29a5ba85dfe2b49b811e96cff57eb38f1d55120a88a1727cfd9fb514ca","observedAt":"2026-10-30T00:32:17.477Z"}}],"participationCoverage":{"total":6,"returned":6,"complete":true,"fullAssertions":"source-metadata-endpoint"},"distribution":null,"publisher":null,"sourceUrl":"https://arxiv.org/abs/2610.36689","snapshotId":"91394d56ae017d2dfb950aa90114624428e1b727781dfba7234195b5a56f5bf5","contentSha256":"06ad2c02a0e0a0b7ca9f9c3948d8210cd96fc1cb67c0c405bfa8d20f765c0f79","textHash":"5ee63d29a5ba85dfe2b49b811e96cff57eb38f1d55120a88a1727cfd9fb514ca","observedAt":"2026-10-30T00:32:17.477Z","basis":"publisher-page-metadata","sourceLinks":44,"linksTruncated":false,"metadataUrl":"/api/pipeline/stream/547090f27774efdc6368996f1ecdd9d5466a4904fde83d8da008169dc13a40b7/source"}},"canonicalUrl":"https://escapemeridian.com/readings/fangyuan-liu-and-coauthors-present-psyjev-for-psychological-547090f2","formats":{"html":"/readings/fangyuan-liu-and-coauthors-present-psyjev-for-psychological-547090f2","markdown":"/readings/fangyuan-liu-and-coauthors-present-psyjev-for-psychological-547090f2.md","json":"/readings/fangyuan-liu-and-coauthors-present-psyjev-for-psychological-547090f2.json"},"attribution":"Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.","maintainedGuideLinks":[],"readerContext":{"version":"reader-context-v1","attribution":"Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.","terms":[{"id":"term-benchmark","name":"Benchmark","explanation":"A defined set of tests used to compare systems. Results depend on the tasks and measures selected.","referenceUrl":"https://arxiv.org/abs/2211.09110"}],"entities":[{"id":"arxiv","name":"arXiv","kind":"organization","context":"arXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.","referenceUrls":["https://arxiv.org/","https://info.arxiv.org/help/api/index.html"],"path":"/entities/arxiv","json":"/entities/arxiv.json","markdown":"/entities/arxiv.md"}]},"researchQuestions":[],"withdrawn":false,"references":{"schemaVersion":1,"target":{"type":"reading","id":"547090f27774efdc6368996f1ecdd9d5466a4904fde83d8da008169dc13a40b7","revision":"5e92e232-898b-4af0-8d35-06e9a9733366","revisionKind":"publication-id","title":"Fangyuan Liu and coauthors present PsyJev for psychological analysis","url":"https://escapemeridian.com/readings/fangyuan-liu-and-coauthors-present-psyjev-for-psychological-547090f2"},"source":{"type":"source","id":"https://arxiv.org/abs/2610.36689","revision":"5ee63d29a5ba85dfe2b49b811e96cff57eb38f1d55120a88a1727cfd9fb514ca","revisionKind":"sha256-source-text","title":"PsyJev: Psychological Analysis via System-1 Judgment","url":"https://arxiv.org/abs/2610.36689"},"status":"available","items":[],"next":null,"scope":"Explicit citations in public readings and public file attachments to this reading. Shared topics are not citations. Other artifact types need an explicit reference adapter."},"revisions":[{"id":"547090f27774efdc6368996f1ecdd9d5466a4904fde83d8da008169dc13a40b7","title":"Fangyuan Liu and coauthors present PsyJev for psychological analysis","argument":"In a paper submitted to arXiv, Fangyuan Liu and coauthors introduce PsyJev, a model for answering questions about psychological information in online self-disclosures. They organize PsyJudgeBench, a benchmark containing varied datasets and questions, and train models to produce answer distributions conditioned on each question. The authors report that their Qwen3.5-based models achieve average test F1 scores of 64.33, 66.21, and 67.56 for the 2B, 4B, and 9B models; they say the 9B model exceeds the strongest external comparison by 2.54 points. Escape Meridian's reading: these results describe performance on the benchmark, not established reliability in real-world psychological or clinical use. How well does PsyJev perform on self-disclosures outside the benchmark, especially when errors could affect people’s care or assessment?","topicIds":["intelligence-and-discovery"],"domain":"science","claimType":"attributed reading","limits":"The evidence is the arXiv abstract; it does not provide the underlying evaluation details or establish how the models perform in real-world psychological or clinical settings.","speaker":{"name":"Liu, Fangyuan, Zhao, Sirui, Wang, Xiaohao, Cheng, Guo, Zhang, Peng, Chen, Enhong","roleAtSource":"Article author"},"source":{"title":"PsyJev: Psychological Analysis via System-1 Judgment","publisher":"arXiv cs.AI","format":"Article","url":"https://arxiv.org/abs/2610.36689","publishedAt":"2026-10-27","dateBasis":"Date basis not provided."},"evidence":[{"text":"The authors introduce PsyJudgeBench, a benchmark with 13 annotated datasets and 82 questions across six psychological areas.","excerpt":"PsyJudgeBench, comprising 13 annotated datasets and 82 questions across six psychological areas","url":"https://arxiv.org/abs/2610.36689","relationship":"Source-stated"},{"text":"The authors introduce PsyJev, which produces answer distributions specified by each request.","excerpt":"We introduce PsyJev, a System-1 model that directly produces distributions over answers specified in each request.","url":"https://arxiv.org/abs/2610.36689","relationship":"Source-stated"},{"text":"The authors report average test F1 scores of 64.33, 66.21, and 67.56 for their Qwen3.5-based 2B, 4B, and 9B models.","excerpt":"The Qwen3.5-based 2B, 4B, and 9B models achieve 64.33, 66.21, and 67.56 average test F1, respectively.","url":"https://arxiv.org/abs/2610.36689","relationship":"Source-stated"},{"text":"The authors report that the 9B model exceeds the strongest external comparison by 2.54 points.","excerpt":"The 9B model exceeds the strongest external comparison by 2.54 points.","url":"https://arxiv.org/abs/2610.36689","relationship":"Source-stated"},{"text":"The abstract reports benchmark test results but does not establish real-world psychological or clinical reliability.","excerpt":"Further analyses examine question-conditioned judgment, inference cost, and probability quality, complemented by qualitative comparisons across models.","url":"https://arxiv.org/abs/2610.36689","relationship":"Escape Meridian interpretation"}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2610.36689","checkedAt":"2026-11-06T15:32:10.001Z","items":[],"note":"No suitable media was found by the supported page extraction."},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"How well does PsyJev perform on self-disclosures outside the benchmark, especially when errors could affect people’s care or assessment?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"Fangyuan Liu and coauthors present PsyJev for psychological analysis","summary":"In a paper submitted to arXiv, Fangyuan Liu and coauthors introduce PsyJev, a model for answering questions about psychological information in online self-disclosures. They organize PsyJudgeBench, a benchmark containing varied datasets and questions, and train models to produce answer distributions conditioned on each question. The authors report that their Qwen3.5-based models achieve average test F1 scores of 64.33, 66.21, and 67.56 for the 2B, 4B, and 9B models; they say the 9B model exceeds the strongest external comparison by 2.54 points. Escape Meridian's reading: these results describe performance on the benchmark, not established reliability in real-world psychological or clinical use. How well does PsyJev perform on self-disclosures outside the benchmark, especially when errors could affect people’s care or assessment?","question":"How well does PsyJev perform on self-disclosures outside the benchmark, especially when errors could affect people’s care or assessment?","claims":[{"text":"The authors introduce PsyJudgeBench, a benchmark with 13 annotated datasets and 82 questions across six psychological areas.","excerpt":"PsyJudgeBench, comprising 13 annotated datasets and 82 questions across six psychological areas","support":"source-stated"},{"text":"The authors introduce PsyJev, which produces answer distributions specified by each request.","excerpt":"We introduce PsyJev, a System-1 model that directly produces distributions over answers specified in each request.","support":"source-stated"},{"text":"The authors report average test F1 scores of 64.33, 66.21, and 67.56 for their Qwen3.5-based 2B, 4B, and 9B models.","excerpt":"The Qwen3.5-based 2B, 4B, and 9B models achieve 64.33, 66.21, and 67.56 average test F1, respectively.","support":"source-stated"},{"text":"The authors report that the 9B model exceeds the strongest external comparison by 2.54 points.","excerpt":"The 9B model exceeds the strongest external comparison by 2.54 points.","support":"source-stated"},{"text":"The abstract reports benchmark test results but does not establish real-world psychological or clinical reliability.","excerpt":"Further analyses examine question-conditioned judgment, inference cost, and probability quality, complemented by qualitative comparisons across models.","support":"interpretation"}],"limitations":["The evidence is the arXiv abstract; it does not provide the underlying evaluation details or establish how the models perform in real-world psychological or clinical settings."],"mediaId":"402625908a648d0345f9eff1194dde1bdfc76ab07adeeff2da6c04edf2211d25","correctionReason":null,"id":"547090f27774efdc6368996f1ecdd9d5466a4904fde83d8da008169dc13a40b7","publicationId":"5e92e232-898b-4af0-8d35-06e9a9733366","version":1,"sourceUrl":"https://arxiv.org/abs/2610.36689","sourceTitle":"PsyJev: Psychological Analysis via System-1 Judgment","publisher":"arXiv cs.AI","publishedAt":null,"originalPublished":{"value":"2026-10-27","precision":"day","basis":"publisher_citation_date"},"firstObservedAt":"2026-10-28T05:28:35.145Z","createdAt":"2026-11-06T11:16:56.477Z","historical":true,"entityIds":["arxiv"],"topicIds":["intelligence-and-discovery"],"media":[{"id":"402625908a648d0345f9eff1194dde1bdfc76ab07adeeff2da6c04edf2211d25","kind":"article","url":"https://arxiv.org/abs/2610.36689","sourceUrl":"https://arxiv.org/abs/2610.36689","label":"Read original","provenance":"source_link","verification":{"status":"verified","checkedAt":"2026-10-30T00:32:18.009Z","method":"acquired_source","contentType":"text/html"},"recommendationReason":"Recommended by Jev for understanding this source; checked by Luna."}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2610.36689","checkedAt":"2026-11-06T15:32:10.001Z","items":[],"note":"No suitable media was found by the supported page extraction."},"editorialDecision":{"outcome":"accept","reasonCodes":["material_development"],"rubricVersion":"stream-admission-v5","mediaId":"402625908a648d0345f9eff1194dde1bdfc76ab07adeeff2da6c04edf2211d25","confidence":0.84,"provider":"typesafe","model":"jev-1.13.0","inputHash":"ef2063d93a0568f6c0c469c52235411296758cf07b27e2235a893f764f6adbce","decidedAt":"2026-10-30T00:32:21.545Z"},"provenance":{"textHash":"5ee63d29a5ba85dfe2b49b811e96cff57eb38f1d55120a88a1727cfd9fb514ca","observationId":"d46f6e99b512644709ddb7adde0ebe971feae8ce145b1fd43748a52a286881b0","retrievedAt":"2026-10-30T00:32:17.477Z","kind":"publisher-page","preparer":{"model":"gpt-6-luna","runId":"f660a074-1687-4610-8955-b741784f4dd6"},"reviewer":{"model":"gpt-6-luna","runId":"3a453453-37d5-4b01-8975-55239cf51dc7"}},"audit":{"status":"uncertain","reviewedAt":"2026-11-06T11:16:56.477Z","model":"anthropic/claude-opus-5.5"},"sourceAttribution":{"authors":[{"name":"Liu, Fangyuan","url":null,"kind":"unknown"},{"name":"Zhao, Sirui","url":null,"kind":"unknown"},{"name":"Wang, Xiaohao","url":null,"kind":"unknown"},{"name":"Cheng, Guo","url":null,"kind":"unknown"},{"name":"Zhang, Peng","url":null,"kind":"unknown"},{"name":"Chen, Enhong","url":null,"kind":"unknown"}],"authorRole":"article-author","hosts":[],"status":"stated","authorRoster":{"truncated":false,"omitted":0,"assertionGroupsOmitted":0},"participation":[{"actor":{"name":"Liu, Fangyuan","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Liu, Fangyuan","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2610.36689","locator":"head/meta[name=author|citation_author]"},"assertionId":"91394d56ae017d2dfb950aa90114624428e1b727781dfba7234195b5a56f5bf5:participation:0","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-10-30T00:32:17.477Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"91394d56ae017d2dfb950aa90114624428e1b727781dfba7234195b5a56f5bf5","sourceUrl":"https://arxiv.org/abs/2610.36689","contentSha256":"06ad2c02a0e0a0b7ca9f9c3948d8210cd96fc1cb67c0c405bfa8d20f765c0f79","textHash":"5ee63d29a5ba85dfe2b49b811e96cff57eb38f1d55120a88a1727cfd9fb514ca","observedAt":"2026-10-30T00:32:17.477Z"}},{"actor":{"name":"Zhao, Sirui","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Zhao, Sirui","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2610.36689","locator":"head/meta[name=author|citation_author]"},"assertionId":"91394d56ae017d2dfb950aa90114624428e1b727781dfba7234195b5a56f5bf5:participation:1","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-10-30T00:32:17.477Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"91394d56ae017d2dfb950aa90114624428e1b727781dfba7234195b5a56f5bf5","sourceUrl":"https://arxiv.org/abs/2610.36689","contentSha256":"06ad2c02a0e0a0b7ca9f9c3948d8210cd96fc1cb67c0c405bfa8d20f765c0f79","textHash":"5ee63d29a5ba85dfe2b49b811e96cff57eb38f1d55120a88a1727cfd9fb514ca","observedAt":"2026-10-30T00:32:17.477Z"}},{"actor":{"name":"Wang, Xiaohao","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Wang, Xiaohao","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2610.36689","locator":"head/meta[name=author|citation_author]"},"assertionId":"91394d56ae017d2dfb950aa90114624428e1b727781dfba7234195b5a56f5bf5:participation:2","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-10-30T00:32:17.477Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"91394d56ae017d2dfb950aa90114624428e1b727781dfba7234195b5a56f5bf5","sourceUrl":"https://arxiv.org/abs/2610.36689","contentSha256":"06ad2c02a0e0a0b7ca9f9c3948d8210cd96fc1cb67c0c405bfa8d20f765c0f79","textHash":"5ee63d29a5ba85dfe2b49b811e96cff57eb38f1d55120a88a1727cfd9fb514ca","observedAt":"2026-10-30T00:32:17.477Z"}},{"actor":{"name":"Cheng, Guo","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Cheng, Guo","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2610.36689","locator":"head/meta[name=author|citation_author]"},"assertionId":"91394d56ae017d2dfb950aa90114624428e1b727781dfba7234195b5a56f5bf5:participation:3","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-10-30T00:32:17.477Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"91394d56ae017d2dfb950aa90114624428e1b727781dfba7234195b5a56f5bf5","sourceUrl":"https://arxiv.org/abs/2610.36689","contentSha256":"06ad2c02a0e0a0b7ca9f9c3948d8210cd96fc1cb67c0c405bfa8d20f765c0f79","textHash":"5ee63d29a5ba85dfe2b49b811e96cff57eb38f1d55120a88a1727cfd9fb514ca","observedAt":"2026-10-30T00:32:17.477Z"}},{"actor":{"name":"Zhang, Peng","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Zhang, Peng","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2610.36689","locator":"head/meta[name=author|citation_author]"},"assertionId":"91394d56ae017d2dfb950aa90114624428e1b727781dfba7234195b5a56f5bf5:participation:4","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-10-30T00:32:17.477Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"91394d56ae017d2dfb950aa90114624428e1b727781dfba7234195b5a56f5bf5","sourceUrl":"https://arxiv.org/abs/2610.36689","contentSha256":"06ad2c02a0e0a0b7ca9f9c3948d8210cd96fc1cb67c0c405bfa8d20f765c0f79","textHash":"5ee63d29a5ba85dfe2b49b811e96cff57eb38f1d55120a88a1727cfd9fb514ca","observedAt":"2026-10-30T00:32:17.477Z"}},{"actor":{"name":"Chen, Enhong","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Chen, Enhong","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2610.36689","locator":"head/meta[name=author|citation_author]"},"assertionId":"91394d56ae017d2dfb950aa90114624428e1b727781dfba7234195b5a56f5bf5:participation:5","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-10-30T00:32:17.477Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"91394d56ae017d2dfb950aa90114624428e1b727781dfba7234195b5a56f5bf5","sourceUrl":"https://arxiv.org/abs/2610.36689","contentSha256":"06ad2c02a0e0a0b7ca9f9c3948d8210cd96fc1cb67c0c405bfa8d20f765c0f79","textHash":"5ee63d29a5ba85dfe2b49b811e96cff57eb38f1d55120a88a1727cfd9fb514ca","observedAt":"2026-10-30T00:32:17.477Z"}}],"participationCoverage":{"total":6,"returned":6,"complete":true,"fullAssertions":"source-metadata-endpoint"},"distribution":null,"publisher":null,"sourceUrl":"https://arxiv.org/abs/2610.36689","snapshotId":"91394d56ae017d2dfb950aa90114624428e1b727781dfba7234195b5a56f5bf5","contentSha256":"06ad2c02a0e0a0b7ca9f9c3948d8210cd96fc1cb67c0c405bfa8d20f765c0f79","textHash":"5ee63d29a5ba85dfe2b49b811e96cff57eb38f1d55120a88a1727cfd9fb514ca","observedAt":"2026-10-30T00:32:17.477Z","basis":"publisher-page-metadata","sourceLinks":44,"linksTruncated":false,"metadataUrl":"/api/pipeline/stream/547090f27774efdc6368996f1ecdd9d5466a4904fde83d8da008169dc13a40b7/source"}}}],"audits":[{"publicationId":"5e92e232-898b-4af0-8d35-06e9a9733366","verdict":"uncertain","findings":[{"dimension":"attribution","status":"pass","explanation":"The draft correctly attributes the paper to Fangyuan Liu and coauthors on arXiv. Results are framed as author reports, and the benchmark-only caveat is labeled as Escape Meridian's reading.","excerpt":"Authors:Fangyuan Liu, Sirui Zhao, Xiaohao Wang, Guo Cheng, Peng Zhang, Enhong Chen"},{"dimension":"summary","status":"uncertain","explanation":"The draft faithfully describes the paper but leaves central terms unexplained for non-specialists. Readers are not told what an F1 score measures, whether these scores are on a 0-100 scale, what 2B/4B/9B (model size in parameters) means, or what it means to produce answer distributions. Without that, the headline results are hard to interpret. Smallest repair: add term-library tooltips or a plain-language gloss for F1, the model sizes and answer distributions. The omission of the JA-RLCD training stage is minor and not a fidelity failure.","excerpt":"The Qwen3.5-based 2B, 4B, and 9B models achieve 64.33, 66.21, and 67.56 average test F1, respectively."},{"dimension":"date","status":"pass","explanation":"originalPublished 2026-10-27 matches the v1 arXiv submission date. The later feed date is correctly not used.","excerpt":"[Submitted on 27 Oct 2026]"},{"dimension":"support","status":"pass","explanation":"The numeric claims, the benchmark composition and the model description match the abstract verbatim. The interpretive caveat that benchmark results do not establish clinical reliability is labeled as interpretation and stays conditional. The question about care and assessment follows plausibly from the source's mention of symptom-relevant information, without asserting harm.","excerpt":"from recognizing expressed experiences to identifying symptom-relevant information"},{"dimension":"media","status":"pass","explanation":"The selected media is the verified direct arXiv abstract page for this paper, which is the most direct access to the source.","excerpt":null}],"createdAt":"2026-11-06T11:16:56.477Z","model":"anthropic/claude-opus-5.5"}],"correctionUrl":"https://escapemeridian.com/corrections/"}