arXiv Computation and Language

Challenges of Auditing: Variability in Outputs of Large Language Models for Health

arXiv Computation and Language
Sep 17

"If I Had to Buy Just ONE: Galaxy S26 Ultra": Auditing AI-Generated Product Recommendations

arXiv:2609.18729v1 Announce Type: cross Abstract: Consumers increasingly use AI chatbots for advice on what to buy. With companies like OpenAI and Google monetising their AI through advertising, this...

By Lucas G. Uberti-Bona Marin, Thales Bertaglia, Giovanni Astante, Bram Rijsbosch, Gijs van Dijck, Anik\'o Hann\'ak, Gerasimos Spanakis, Konrad Kollnig
arXiv Computation and Language
Oct 23

SIM-UT: A User Simulation based Evaluation Framework for Medical Triage Chatbots

SIM-UT is a multi‑turn, user‑simulation evaluation framework designed to uncover realistic failure modes in large‑language‑model (LLM) medical triage chatbots. The study applied SIM‑UT to three WhatsApp‑based health chatbots in India, South Africa, and Kenya, comparing a medically fine‑tuned model (MedGemma) with general‑purpose models (ChatGPT and Claude) across more than 5,000 multilingual simulated conversations. A prominent failure identified was de‑escalation risk, where chatbots incorrectly down‑rated serious symptoms after users minimized them, potentially leading to dangerous clinical outcomes.

By Deepali Mittal, Dheeraj Mundhra, Yulia Sidorova, Tersia de Kock, Oli Munns, Sofie S Meyer
arXiv AI
Nov 5

TherapyGym: Evaluating and Aligning Clinical Fidelity and Safety in Therapy Chatbots

TherapyGym is a framework designed to evaluate and improve therapy chatbots on two key clinical dimensions: fidelity to evidence-based practice and safety. It uses the Cognitive Therapy Rating Scale (CTRS) to automatically score adherence to CBT techniques across multi‑turn sessions, and a multi‑label annotation scheme to assess therapy‑specific risks such as failure to address harm or abuse. The framework also introduces THERAPYJUDGEBENCH, a validation set of 116 dialogues with 1,270 expert ratings, and serves as a training harness where CTRS and safety rewards guide reinforcement learning with diverse patient simulations, leading to significant improvements in expert‑rated performance.

By Fangrui Huang, Souhad Chbeir, Arpandeep Khatua, Sheng Wang, Sijun Tan, Kenan Ye, Lily Bailey, Merryn Daniel, Ryan Louie, Sanmi Koyejo, Ehsan Adeli
arXiv Machine Learning
Jul 17

Addressing Benchmarking Gaps in Large Language Models for Health and Medicine with Dynamic Red-Teaming

arXiv:2508. 00923v3 Announce Type: replace Abstract: Large language models (LLMs) are increasingly used to answer health-related questions and support healthcare workflows, yet evidence for their safety still relies heavily on static benchmarks that can rapidly become obsolete or be optimized against.

By Jiazhen Pan (Cherise), Bailiang Jian (Cherise), Paul Hager (Cherise), Yundi Zhang (Cherise), Che Liu (Cherise), Friederike Jungmann (Cherise), Hongwei Bran Li (Cherise), Julian Canisius (Cherise), Chenyu You (Cherise), Junde Wu (Cherise), Jiayuan Zhu (Cherise), Fenglin Liu (Cherise), Yuyuan Liu (Cherise), Niklas Bubeck (Cherise), Moritz Knolle (Cherise), Chen (Cherise), Chen (Cherise), Christian Wachinger, Zhenyu Gong, Cheng Ouyang, Georgios Kaissis, Benedikt Wiestler, Daniel Rueckert
arXiv AI
Sep 15

A primer on evaluation methods for large language models in healthcare

arXiv:2609.14819v1 Announce Type: cross Abstract: Large language models (LLMs) have a growing range of applications in medicine, and their evaluation is critical for ensuring they provide benefit and...

By Suzannah E McKinney, Phuc Vu, Samuel A Justice, Christopher Humphries, Alyssa Pradhan, Timothy J Keyes, Bernardo C Bizzo, Keith J Dreyer, Sarah F Mercaldo, James M Hillis
arXiv AI
Oct 31

PromptSense: Sentiment-Aware Mental Health Chatbot with Dynamic Prompting

PromptSense introduces a sentiment‑aware mental health chatbot that incorporates sentiment and depression analysis with rule‑based filtering before model inference. By dynamically constructing prompts from emotional and risk insights, the system enhances contextual relevance, empathy, and safety of responses. Experiments across four model strategies and two prompt variants show that the engineered prompts outperform DSFT and RLAIF by 1.8% and 1.1% on qualitative metrics and toxicity assessments.

By Sugi Lee, Jianlong Zhou, Jing sUN
Hugging Face Trending Papers
Sep 8

API Benchmark Scores Do Not Reliably Transfer to Chatbot Interfaces

The paper examines whether benchmark scores obtained via APIs accurately reflect the performance of AI chatbots when accessed through user interfaces. By auditing ChatGPT, Claude, and Gemini across seven systems and nine benchmarks, the authors find that API evaluations consistently overestimate accuracy and consistency compared to interface evaluations, with differences comparable to downgrading a full model generation. Attempts to align API behavior with interface behavior through prompt and parameter adjustments only partially close the gap.