---
title: "PsyJev: Psychological Analysis via System-1 Judgment"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2610.36689v1"
permalink: "https://feed.astralor.com/items/itm_1576badc0b19bc424d50986b"
published: "2026-10-27T06:03:37.000Z"
score: 56
channel: "tech"
tags: ["推理效率","RAG","多模态","PsyJev","Psychological Analysis","System-1 Judgment Self-disclosures","Web","Self-disclosures"]
---

# PsyJev: Psychological Analysis via System-1 Judgment

## 速览

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…

## 关键点

- 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.

## 全文

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.

---

引用前请核对原文;本文由星际流动整理:https://feed.astralor.com/items/itm_1576badc0b19bc424d50986b