arXiv Artificial IntelligenceAgentDrift: A Step-Labeled Benchmark of Injection-Hijacked LLM Agent Trajectories
arXiv:2609.06972v1 Announce Type: cross Abstract: LLM agents complete tasks by issuing sequences of tool…
Curated from international AI laboratories, specialist publications and technology outlets. Last update: 3 hours ago.
arXiv Artificial IntelligencearXiv:2609.06972v1 Announce Type: cross Abstract: LLM agents complete tasks by issuing sequences of tool…
arXiv Artificial IntelligencearXiv:2609.06967v1 Announce Type: cross Abstract: Ensuring effective transfer learning for vision-language…
arXiv Artificial IntelligencearXiv:2609.06962v1 Announce Type: cross Abstract: Real-world remote sensing image dehazing (RSID) remains…
arXiv Artificial IntelligencearXiv:2609.06959v1 Announce Type: cross Abstract: 3D point cloud semantic segmentation is essential for…
arXiv Artificial IntelligencearXiv:2609.06958v1 Announce Type: cross Abstract: Reinforcement learning has become the leading paradigm in…
arXiv Artificial IntelligencearXiv:2609.06951v1 Announce Type: cross Abstract: Activation steering promises modular control of language…
arXiv Artificial IntelligencearXiv:2609.06942v1 Announce Type: cross Abstract: Medium-range precipitation forecasts are impaired by…
arXiv Artificial IntelligencearXiv:2609.06934v1 Announce Type: cross Abstract: Post-hoc safety training (RLHF, DPO) is the dominant way to…
arXiv Artificial IntelligencearXiv:2609.06928v1 Announce Type: cross Abstract: Emotion analysis is a fundamental task in computer vision…
arXiv Artificial IntelligencearXiv:2609.06921v1 Announce Type: cross Abstract: We study constrained online convex optimization with…
arXiv Artificial IntelligencearXiv:2609.06912v1 Announce Type: cross Abstract: Tabular foundation models (TFMs) are commonly pretrained on…
arXiv Artificial IntelligencearXiv:2609.06887v1 Announce Type: cross Abstract: Automated experimentation is most effective when the…