This study demonstrates that an LLM's assigned support role (e.g., Inform, Coach, Relate) significantly alters its safety profile and the types of risks it presents when assisting users in complex car…

This paper systematically evaluates how LLMs uncritically adapt to potentially dangerous user prompts related to eating disorders, finding that specific linguistic cues significantly increase the like…

The paper introduces LLUMI, an open-source framework that improves LLM writing assistance for mental health support using community feedback, demonstrating comparable performance to proprietary models…

The paper argues that LLM guardrails and persona dynamics create an unethical 'reality gap' by laundering epistemic risk onto users, advocating for task-level causal requirements over response-level m…

The paper introduces Responsible Contrastive Soft Prompting (RCSP), a parameter-efficient method using soft prompts to improve LLM reliability by simultaneously suppressing hallucinations, encouraging…

This study investigated user reactions to inferred personal information from their own ChatGPT histories, finding that acceptability is governed by context-sensitive norms regarding generation, retent…

This paper introduces Ghostwriter, an attack framework demonstrating that LLMs are highly vulnerable to adopting misleading viewpoints when provided with fabricated, yet credible-looking, evidence.

The paper demonstrates that increasing the toxicity of prompts significantly degrades the factual reliability of LLMs, a degradation linked to the selective amplification of perturbation-sensitive nod…

This paper introduces HarmAmp, a new benchmark for multi-turn harm amplification, and proposes TrajSafe, a proactive monitoring system that significantly reduces harmfulness in LLM interactions while…

The paper introduces EUDAIMONIA, a new framework and benchmark for evaluating how well LLMs align with user welfare in social interactions, finding that even state-of-the-art models frequently violate…

The paper introduces Med-HEAL, a comprehensive framework and dataset for systematically identifying and mitigating hallucinations in medical LLMs, demonstrating that a self-critique pipeline significa…

The paper introduces the DECK taxonomy, a novel framework that classifies LLM hallucinations not by their content error, but by their detectability signature based on inter-sample consistency and toke…

The paper analyzes that while multimodal large language models (MLLMs) offer superior semantic understanding for image generation, this enhanced capability significantly increases safety risks, partic…

The paper proposes 'Think Fast, Talk Smart,' a pipeline that separates deterministic data analysis from LLM generation, showing that offloading recurring, structured tasks to code significantly improv…

The paper introduces SuiChat-CN, a novel Chinese group-chat benchmark for contextual suicide risk assessment, demonstrating that multi-party conversational context is crucial for accurate detection.

This paper introduces the concept of Safety Geometry Collapse, demonstrating that multimodal inputs degrade the safety separation of LLMs, and proposes ReGap, a training-free method that adaptively co…

This paper evaluates multiple LLMs (DeepSeek-R1, OpenBioLLM-Llama3, Qwen 3.5) for generating privacy-safe, high-quality synthetic mental health reports, demonstrating their effectiveness in expanding…

This paper introduces FreakOut-LLM, demonstrating that emotional context, specifically stress, significantly compromises the safety alignment of large language models, increasing jailbreak susceptibil…

The paper evaluates the semantic stability of clinical LLMs to linguistic variations, finding that domain specialization does not guarantee consistent robustness improvements.

RealityTest introduces a large-scale, multimodal, and multilingual benchmark using real-world human data to test how AI systems disclose their identity, finding that context and phrasing are more crit…