The paper introduces RAISE, a diagnostic framework that tests whether a costly large language model (LLM) signal provides enough pre-call information to justify selective use. It identifies the failure mode of acquisition collapse, where an LLM appears useful overall but lacks actionable evidence for individual decisions. The authors demonstrate RAISE with Structured Hypothesis Embeddings (SHE) and evaluate it across multiple study designs, showing that predictable incremental benefit, rather than average lift, indicates recoverable selective value.
By Ying Yuan, Yu Wang, Yize Cheng, Xuyang Wu
arXiv:2606. 10747v1 Announce Type: new Abstract: As AI systems built from multiple language-model agents become more common, they are increasingly used to make decisions together: discussing, negotiating, and acting on shared tasks.
By Filippo Tonini, Federico Torrielli, Anton Danholt Lautrup, Peter Schneider-Kamp, Mustafa Mert \c{C}elikok, Lukas Galke Poech
The paper introduces a deterministic, reproducible e‑commerce environment that pre‑commits customer and trajectory parameters, enabling a simulated consumer to attempt purchasing a target cart with the help of an evaluated model. The environment records every assistant action and state, allowing post‑trial evaluation of specific conversation components and applying penalties based on tool‑call accuracy. Using this setup, the authors benchmark eight open‑weight agents (20B–35B parameters) across 160 trials and 44 metrics, revealing nuanced performance issues such as under‑action, over‑purchase, unsupported product attributes, and poor search that are hidden by overall success rates.
By Nimit Shah, Haitz S\'aez de Oc\'ariz Borde
The paper introduces ToxicBench, a benchmark designed to evaluate how tool‑augmented data agents handle incorrect tool outputs. By pairing clean and poisoned observations across numerical, label, schema, and retrieval errors, the authors assess both the checking process and the final answer adoption. In a 118‑task GPT evaluation, poisoning reduces task success by 26–39 percentage points, revealing that repeated poisoning leads to wrong-answer adoption even after checking, while ordinary retries help only under one‑shot poisoning. Human annotations on 200 trajectories confirm the scoring system’s reliability, showing 96% agreement with task success and supporting the benefits of retries and audit‑based adoption.
By Zifu Tao, Changqing Yin
arXiv:2609.35889v1 Announce Type: cross
Abstract: Tool-using language-model agents select and execute third-party artifacts. Different implementations can return the requested output while producing...
By Xiaoyu Xu, Zi Liang, Minxin Du, Qipeng Xie, Qingqing Ye, Yuyuan Li, Haibo Hu
arXiv:2608. 05235v1 Announce Type: cross Abstract: Research agents increasingly conduct multi-round machine-learning experiments in industrial recommendation settings and retain the resulting trajectories to guide later decisions.
By Zijie Zhuang, Changxin Lao, Pengbo Xu, Hanwen Xu, Ruochen Yang, Yingzhi He, Peng Zhang, Jiangxia Cao, Yusheng Huang, Guohong Mu, Jian Liang, Ruiming Tang, Shuang Yang, Zhaojie Liu, Wenwu Ou, Kun Gai
arXiv:2607. 18867v1 Announce Type: new Abstract: Large language models leak parametric knowledge of realized outcomes into historical financial decision tasks.
By Haozhe Jia
arXiv:2608.30685v1 Announce Type: new
Abstract: Large language model (LLM) agents are increasingly deployed in user-facing services that require iterative tool use under dynamic business conditions....
By Wei Chen, Peilun Zhou, Zhaoyu Hu, Jiajun Chai, Zhongni Hou, Yufei Zhang, Derong Xu, Guojun Yin, Wei Lin, Zhi Zheng, Tong Xu
The paper introduces a framework for evaluating AI systems that not only checks final labels but also tracks the reasoning behind them through three core sources—grounds, norms, and authority—forming an eight-cell counterfactual judgment cube. It defines minimal source replacement sets, called judgment receipts, to explain changes in verdicts and provides certification cost bounds for black-box evaluators. The authors present ReasonBench, a benchmark with 19,520 cases, and demonstrate that while high standard accuracy can mask robustness issues, receipt accuracy reveals significant gaps in reasoning consistency across different models.
By Ye Chen, Weining Zhang
The paper presents a layered framework for evaluating conversational AI by aligning offline proxy signals with online A/B experiment outcomes. It introduces a three‑step alignment chain—behavioral label to product outcome, classifier to candidate behavior, and offline signal to experiment effect—alongside an audit protocol that compares confidence intervals and rankings. In a real‑world deployment, the composite proxy achieved 81.1% F1 versus 34.3% for the raw classifier, correctly predicting direction on all 113 contrasts and enabling efficient prioritization of candidate models before costly online testing.
By Xuanyi Li, Vaskar Nath, Hossein Amirkhani, Jay Li, Alex Deng
arXiv:2606. 10794v3 Announce Type: replace Abstract: Existing black-box LLM provenance methods achieve comparability by querying every candidate model with the same diagnostic prompts.
By Jiaxu Liu, Sunnan Mu, Dong Huang, Liuyin Wang, Jing Shao, Jie Zhang
The paper audits a commercial non‑generative System‑1 model on biosecurity‑relevant benchmarks, evaluating accuracy, calibration, error detection, selective prediction, and sensitivity to answer‑option order. It finds that the model’s accuracy varies strongly by task, is reasonably well calibrated when the vendor’s uncertainty field is interpreted correctly, and that option order can cause significant prediction changes—averaging across rotations improves accuracy. The study also shows that applying averaging only to low‑confidence items recovers most of the gain at a lower cost.
By Kimon Antonios Provatas, Ilias Georgakopoulos-Soares