{"ID":2894221,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2507.10958","arxiv_id":"2507.10958","title":"DS@GT at eRisk 2025: From prompts to predictions, benchmarking early depression detection with conversational agent based assessments and temporal attention models","abstract":"This Working Note summarizes the participation of the DS@GT team in two eRisk 2025 challenges. For the Pilot Task on conversational depression detection with large language-models (LLMs), we adopted a prompt-engineering strategy in which diverse LLMs conducted BDI-II-based assessments and produced structured JSON outputs. Because ground-truth labels were unavailable, we evaluated cross-model agreement and internal consistency. Our prompt design methodology aligned model outputs with BDI-II criteria and enabled the analysis of conversational cues that influenced the prediction of symptoms. Our best submission, second on the official leaderboard, achieved DCHR = 0.50, ADODL = 0.89, and ASHR = 0.27.","short_abstract":"This Working Note summarizes the participation of the DS@GT team in two eRisk 2025 challenges. For the Pilot Task on conversational depression detection with large language-models (LLMs), we adopted a prompt-engineering strategy in which diverse LLMs conducted BDI-II-based assessments and produced structured JSON outpu...","url_abs":"https://arxiv.org/abs/2507.10958","url_pdf":"https://arxiv.org/pdf/2507.10958v1","authors":"[\"Anthony Miyaguchi\",\"David Guecha\",\"Yuwen Chiu\",\"Sidharth Gaur\"]","published":"2025-07-15T03:40:46Z","proceeding":"cs.CL","tasks":"[\"cs.CL\"]","methods":"[\"Large Language Model\"]","has_code":false}