arXiv:2606. 11063v1 Announce Type: new Abstract: AI control protocols oversee untrusted models by monitoring their actions and modifying potentially unsafe steps, often using a trusted model.
By Joachim Schaeffer, Thomas Jiralerspong, Alexander Panfilov, Guillaume Lajoie, Jonas Geiping, Yoshua Bengio, Roland S. Zimmermann
arXiv:2603. 22016v3 Announce Type: replace-cross Abstract: Large Reasoning Models (LRMs) often reach a correct solution before their long Chain-of-Thought trace ends, yet continue with redundant verification, repeated attempts, or unnecessary exploration that wastes computation and can even overturn the correct answer.
By Xinyan Wang, Xiaogeng Liu, Ming Pei, Chaowei Xiao
The paper introduces a live trace model that incrementally folds an append‑only event ledger into typed run state, producing per‑consumer views for both human observers and the agent itself. Evaluations show that for observers, the compiled view reduces input tokens by 14–15× and cost by 5–7× while improving accuracy from 0.48 to 0.85–0.87. For agents, maintaining running statistics in per‑step state enables success on 120‑link sequential tasks where full‑context prompting fails, and a prompt‑level scratchpad matches the fold’s accuracy at lower cost.
By Egor Pakhomov, Erik Nijkamp
arXiv:2604. 02478v2 Announce Type: replace Abstract: Deep learning models excel at detecting anomaly patterns in normal data.
By Jiyong Kwon, Ujin Jeon, Sooji Lee, Guang Lin
arXiv:2606. 04296v1 Announce Type: new Abstract: As autonomous AI agents move from conversational systems to long-horizon software execution, runtime safety layers that decide when to interrupt an agent have become essential.
By Manvendra Modgil
ObserverBench is a benchmark framework that evaluates whether internal mechanistic estimators—called observers—are suitable for guiding interventions, control, or safety actions in language models. It separates estimation accuracy from the loss incurred by the chosen action, showing that accurate predictions do not always lead to better decisions. Experiments on GPT‑2‑small, Qwen2.5‑7B, Gemma‑2‑9B‑it, and Qwen3.5‑9B demonstrate that observers trained on action loss can reduce deployment loss, while traditional metrics like AUROC may rank monitors differently from actual performance.
By Vijay Erramilli