Automated agent design increasingly trusts a single score to decide whether a change is an improvement -- but a score optimized aggressively enough ceases to measure what it was meant to. When an agent also explains its design, that explanation is a second signal we can check without human labels: negate each stated claim, re-run, and observe whether performance drops. We ask whether such a verifier stays reliable once an agent optimizes against it. Our contribution is not a new verifier but a way to stress-test one: a pressure ratio $ρ$ measuring how much the verifier steers the agent's search, and four falsifiable probes covering four failure modes -- giving no signal, being too weak to matter, being gamed by stating fewer claims, and being queried before the agent commits. We apply the test to our own verifier, the Rationale Consistency Score (RCS), inside an LLM agent that designs machine-learning pipelines, with one generator model, six tasks, and three seeds. First, in our prior configuration the verifier barely affects the search objective ($ρ\ll 1$), partly because of an accuracy gate, so its robustness had not been tested; it still shapes the reported artifact through a final-selection tie-break that $ρ$ does not register. Second, a capability control shows the agent shortens its claim lists when instructed, yet search does not, even at $ρ$ over 100 times the prior level; since feedback omits claim volume and horizons are short, this null result is largely structural rather than evidence of robustness. Third, when the agent can probe the verifier before committing, it keeps claims that pass while held-out faithfulness falls -- a decline invisible to in-loop metrics, though partly confounded by how RCS scores small pipelines. We offer these as preliminary evidence and propose a reporting standard: audit a verifier with $ρ$ and these probes before trusting it.
CheXPO: Preference Optimization for Chest X-ray VLMs with Counterfactual Rationale
The Key Laboratory of Smart Human-Computer Interaction(智能人机交互关键实验室);Wearable Technology of Shaanxi Province, Xidian University, Xi'an, China(陕西 wearable 技术, 西安电子科技大学, 中国);Nanjing University of Science and Technology(南京理工大学)
Reading the Mask: Descriptor-Grounded BI-RADS Rationale Generation for Breast MRI Segmentation
Indian Institute of Technology Roorkee(鲁尔基印度理工学院);Indian Institute of Technology Madras(马德拉斯印度理工学院);Mehta Family School of Data Science and Artificial Intelligence, Indian Institute of Technology Roorkee(鲁尔基印度理工学院梅塔家族数据科学与人工智能学院)
Causal Consistency Regularization: Training Verifiably Sensitive Reasoning in Large Language Models
Department of Computer Science, Iowa State University(计算机科学系,爱荷华州立大学);Department of Civil, Construction & Environmental Engineering, Iowa State University(土木、建筑与环境工程系,爱荷华州立大学)
A rubric landscape for evaluating clinical reasoning in large language models: what exists, what is missing, and what needs to be combined
DRIVE-Health CDT(DRIVE-Health 博士训练中心);Department of Biostatistics and Health Informatics(生物统计学与健康信息学系);Institute of Psychiatry, Psychology and Neuroscience(精神病学、心理学与神经科学研究所);King’s College London(伦敦国王学院);Cleveland Clinic London(伦敦克利夫兰诊所)