Scholarlink

<p>Large-scale online mental health services increasingly use language models to detect suicide risk in chat conversations, yet clinicians often lack transparent, clinically relevant explanations for model decisions. We propose a dual-mode framework combining extractive and abstractive layers: highlighting help-seeker utterances most responsible for predictions, and mapping them to psychological risk categories from a Suicide Risk Factors (SRF) lexicon. We apply this to thousands of Hebrew hotline chat sessions and English Reddit posts. For extraction, our BCombined method integrates SHAP, LIME, Integrated Gradients, and Embedding-shift relevance via Borda voting, outperforming baselines in sufficiency, completeness, and predictive power. For abstraction, a Llama-3.1-based layer improves alignment with expert SRF annotations on F1 and Intersection over Union (IoU). In a user study with hotline counselors, the abstractive layer improved perceived helpfulness and understandability without increasing cognitive load. Our findings support explainable NLP in high-stakes clinical workflows, demonstrating generalizability across two languages and bridging the gap between model output and psychological reasoning.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Dual-Mode Explanations for Suicide Risk Detection in Online Mental Health Support

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月25日

查看全文
  • Noam Munz,
  • Avi Segal,
  • Meytal Grimland,
  • Hadas Yeshayahu,
  • Joy Benatov,
  • Inbar Shenfeld,
  • Loona Ben Dayan,
  • Yossi Levi-Belz,
  • Kobi Gal

摘要

Large-scale online mental health services increasingly use language models to detect suicide risk in chat conversations, yet clinicians often lack transparent, clinically relevant explanations for model decisions. We propose a dual-mode framework combining extractive and abstractive layers: highlighting help-seeker utterances most responsible for predictions, and mapping them to psychological risk categories from a Suicide Risk Factors (SRF) lexicon. We apply this to thousands of Hebrew hotline chat sessions and English Reddit posts. For extraction, our BCombined method integrates SHAP, LIME, Integrated Gradients, and Embedding-shift relevance via Borda voting, outperforming baselines in sufficiency, completeness, and predictive power. For abstraction, a Llama-3.1-based layer improves alignment with expert SRF annotations on F1 and Intersection over Union (IoU). In a user study with hotline counselors, the abstractive layer improved perceived helpfulness and understandability without increasing cognitive load. Our findings support explainable NLP in high-stakes clinical workflows, demonstrating generalizability across two languages and bridging the gap between model output and psychological reasoning.