arXiv cs.AI· Fangyuan Liu, Sirui Zhao, Xiaohao Wang, Guo Cheng, Peng Zhang, Enhong Chen·· 3 小时前AI 评分34
PsyJev：基于 System-1 判断的心理分析模型
PsyJev: Psychological Analysis via System-1 Judgment
AI 导读
arXiv 论文提出 PsyJev，一种直接输出答案分布的 System-1 心理分析模型，并构建含 13 个标注数据集、82 个问题、覆盖六个心理领域的 PsyJudgeBench。
正文 · 原文
Abstract:Self-disclosures on the Web support diverse psychological analyses, from recognizing expressed experiences to identifying symptom-relevant information. We study a shared, question-conditioned decision interface for this heterogeneous problem in content analysis. We organize PsyJudgeBench, comprising 13 annotated datasets and 82 questions across six psychological areas, while preserving source-specific criteria and masks for unobserved labels. We introduce PsyJev, a System-1 model that directly produces distributions over answers specified in each request. Joint supervised training establishes the shared judgment function, followed by Judgment-Anchored RLCD (JA-RLCD), which adapts logit-space reward optimization to heterogeneous psychological judgments through balanced supervision, selective retention of correct SFT predictions, and question-group reward optimization. The Qwen3.5-based 2B, 4B, and 9B models achieve 64.33, 66.21, and 67.56 average test F1, respectively. The 9B model exceeds the strongest external comparison by 2.54 points. Controlled 2B comparisons examine the training stages and anchored refinement. Further analyses examine question-conditioned judgment, inference cost, and probability quality, complemented by qualitative comparisons across models.
| Comments: | 12 pages, 5 figures, 14 tables |
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.36689 [cs.AI] |
| (or arXiv:2610.36689v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.36689 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Fangyuan Liu [view email]
[v1]
Tue, 27 Oct 2026 06:03:37 UTC (1,094 KB)
来源：arXiv cs.AI · arxiv.org