CUNY at CLPsych 2026: A Pipeline Approach to Classification and Summarization of Mental Health Changes
Quick Answer
This paper shows that CUNY's submission to CLPsych 2026 utilized ensemble in-context learning from three large language models, achieving first place in Task 1.1 and notable rankings in other tasks.
Quick Take
CUNY's submission to CLPsych 2026 utilized ensemble in-context learning from three large language models, achieving first place in Task 1.1 and notable rankings in other tasks. The approach effectively captured mental health changes through social media, demonstrating significant performance improvements over baseline methods.
Key Points
- Employed ensemble learning from three open-weight large language models for mental health classification.
- Ranked first in Task 1.1, fourth in Tasks 1.2 and 2, and third in Task 3.1.
- Utilized majority voting to infer dominant self-states from social media posts.
- Augmented in-context example labels to improve mood dynamics summarization.
- Source code for the experiments is publicly available on GitHub.
Paper Resources
Article Excerpt
From source RSS / original summaryarXiv:2605. 24164v1 Announce Type: new Abstract: We describe our submission to the CLPsych~2026 Shared Task on capturing and characterizing mental health changes through social media timeline dynamics. To infer the dominant self-states in posts (Tasks 1. 1 and 1. 2), we ensemble in-context learning of three open-weight large language models using majority voting. For predicting moments of change in a timeline (Task~2), we train supervised classifiers on features derived from Task~1. 1 predictions.
To summarize the patterns of mood dynamics and their progression over time within a timeline (Task 3. 1), we augment in-context example labels predicted by upstream systems (Tasks 1. 1, 1. 2, and 2), yielding performance gains over zero-shot and unaugmented in-context learning baselines. Our submission ranked first on Task~1. 1, fourth on Task~1. 2, fourth on Task~2, and third on Task~3. 1. \footnote{The source code for the experiments is available at https://github. com/amirzia/clpsych26-cuny
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.CL
See more →Quantifying Prior Dominance in Systems
The study introduces the Normalized Context Utilization (NCU) metric to evaluate Retrieval-Augmented Generation (RAG) systems, revealing that Small Language Models (SLMs) outperform larger models in factual extraction. The findings indicate that traditional scaling laws yield diminishing returns, with a commercial API frequently failing against adversarial evidence due to systemic confidence collapse.