aiqbench analysis
This research introduces CRPO, a reinforcement learning framework that aligns large language model reasoning with human cognitive appraisal processes for mental health assessment. By incorporating stage-dependent uncertainty modeling and entropy regularization, the method improves weighted F1-scores by over 10 percentage points across eight datasets while enhancing interpretability in high-stakes scenarios like suicide risk evaluation. The work demonstrates a significant advancement in tailoring AI models to specific clinical domains through theory-grounded optimization techniques rather than generic post-training approaches.
Mental-R1: Aligning LLM Reasoning for Mental Health Assessment
arXiv:2606.13176v1 Announce Type: new Abstract: Mental health problems such as anxiety, depression, and suicide remain urgent global challenges, where timely and accurate assessment is critical for effective intervention. Recently, large language models have been explored for mental health assessment. However, existing general-purpose post-training methods do not align with the cognitive processes of human assessment, which may lead to unreliable reasoning outcomes. To bridge this gap, we propose Cognitive Relative Policy Optimization (CRPO), a reinforcement learning framework tailored for the mental health domain. CRPO extends group relative policy optimization by integrating stage-dependent uncertainty modeling into the policy optimization process. Specifically, we introduce a stage-wise entropy regularization mechanism that encourages broad exploration in early reasoning phases and progressively enforces confident decision-making in later stages, mimicking the human cognitive shift from uncertainty to certainty. In addition, inspired by cognitive appraisal theory, we formalize cognitive reasoning stages, thereby guiding theory-grounded interpretable inference. Experiments on 8 mental health datasets show that CRPO achieves an average improvement of 10.4 percentage points in weighted F1-score over the best reinforcement learning baseline. Furthermore, the CRPO-trained model Mental-R1 demonstrates clear advantages compared with existing large language models on reasoning-intensive cases, suggesting that CRPO enhances reasoning capabilities for mental health assessment.
Abstract:Mental health problems such as anxiety, depression, and suicide remain urgent global challenges, where timely and accurate assessment is critical for effective intervention. Recently, large language models have been explored for mental health assessment. However, existing general-purpose post-training methods do not align with the cognitive processes of human assessment, which may lead to unreliable reasoning outcomes. To bridge this gap, we propose Cognitive Relative Policy Optimization (CRPO), a reinforcement learning framework tailored for the mental health domain. CRPO extends group relative policy optimization by integrating stage-dependent uncertainty modeling into the policy optimization process. Specifically, we introduce a stage-wise entropy regularization mechanism that encourages broad exploration in early reasoning phases and progressively enforces confident decision-making in later stages, mimicking the human cognitive shift from uncertainty to certainty. In addition, inspired by cognitive appraisal theory, we formalize cognitive reasoning stages, thereby guiding theory-grounded interpretable inference. Experiments on 8 mental health datasets show that CRPO achieves an average improvement of 10.4 percentage points in weighted F1-score over the best reinforcement learning baseline. Furthermore, the CRPO-trained model Mental-R1 demonstrates clear advantages compared with existing large language models on reasoning-intensive cases, suggesting that CRPO enhances reasoning capabilities for mental health assessment.
Artificial Intelligence (cs.AI)
arXiv:2606.13176 [cs.AI]
(or arXiv:2606.13176v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2606.13176 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 290 words
Related coverage
Hint-based CoT faithfulness evals still mostly work on Claude
LessWrong · Jul 30, 2026
Mental-R1: Aligning LLM Reasoning for Mental Health Assessment
arXiv · Oct 6, 2026
Construct validity of Claude Opus 4.8's System Card – A commentary
LessWrong · Jun 12, 2026
The J-Space Debate, Agent Swarms, and Pacing Frontier AI - Digital Minds Newsletter #4
LessWrong · Sep 18, 2026