New PsyJev Model Enhances Psychological Analysis with Qwen3.5 · arXiv research

Researchers have introduced PsyJev, a novel System-1 model designed for psychological analysis of web-based self-disclosures. This model is trained on PsyJudgeBench, a comprehensive dataset comprising 13 annotated datasets and 82 questions across six psychological domains. PsyJev utilizes a joint supervised training approach followed by Judgment-Anchored RLCD to optimize its judgment function for heterogeneous psychological assessments. The Qwen3.5-based models demonstrated strong performance, with the 9B version achieving an average test F1 score of 67.56, surpassing external comparisons by 2.54 points. AI

IMPACT This research introduces a new method for analyzing psychological states from text, potentially improving AI's ability to understand and respond to human emotional and mental health needs.

RANK_REASON The cluster describes a new research paper detailing a novel model and dataset for psychological analysis. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• arXiv
• JA-RLCD
• PsyJev
• PsyJudgeBench
• Qwen 3.5

• paper
• model release
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →