Research radar

arXiv preprints, filtered for developer impact.

Concise summaries of new AI papers, ranked for how likely they are to change how production systems are built. arXiv collection remains rate-limited.

Papers tracked

180

1 arXiv sources

Top radar

http://arxiv.org/abs/2605.30352v1

Published 24h

by paper date

Rate limit

1/3s

arXiv policy

Paper	Authors	arXiv ID	Categories	Published	Code	Radar	Summary
Reasoning with Sampling: Cutting at Decision Points	Felix Zhou, Anay Mehrotra, Quanquan C. Liu	http://arxiv.org/abs/2605.30327v1	cs.LG, cs.AI	May 28, 2026	None	RDR85	Frontier reasoning models are produced by posttraining base language models with reinforcemen...
SpatialBench: Is Your Spatial Foundation Model an All-Round Player?	Haosong Peng, Hao Li, Jiaqi Chen	http://arxiv.org/abs/2605.27367v1	cs.CV	May 26, 2026	None	RDR85	While spatial foundation models have demonstrated impressive performance on standard datasets...
MobileMoE: Scaling On-Device Mixture of Experts	Yanbei Chen, Hanxian Huang, Ernie Chang	http://arxiv.org/abs/2605.27358v1	cs.LG, cs.AI	May 26, 2026	None	RDR85	Mixture-of-Experts (MoE) has become the de facto architecture for hundred-billion-parameter l...
LLMSurgeon: Diagnosing Data Mixture of Large Language Models	Yaxin Luo, Jiacheng Cui, Xiaohan Zhao	http://arxiv.org/abs/2605.30348v1	cs.CL, cs.AI	May 28, 2026	None	RDR84	The pretraining data mixture of Large Language Models (LLMs) constitutes their "digital DNA",...
Bias Leaves a Gradient Trail: Label-Free Bias Identification via Gradient Probes on Concept Decompositions	Thomas Vitry, Kieran Edgeworth, Stefan Wermter	http://arxiv.org/abs/2605.28780v1	cs.CV, cs.LG	May 27, 2026	Detected	RDR84	Vision classifiers can exploit spurious correlations, achieving high in-distribution accuracy...
Automated Benchmark Auditing for AI Agents and Large Language Models	Junlin Wang, Federico Bianchi, Shang Zhu	http://arxiv.org/abs/2605.26079v1	cs.CL	May 25, 2026	None	RDR84	Modern AI benchmarks operate at a complexity that outpaces traditional verification methods....
SkillOpt: Executive Strategy for Self-Evolving Agent Skills	Yifan Yang, Ziyang Gong, Weiquan Huang	http://arxiv.org/abs/2605.23904v1	cs.AI, cs.CL	May 22, 2026	None	RDR84	Agent skills today are hand-crafted, generated one-shot, or evolved through loosely controlle...
Decomposing Queries into Tool Calls for Long-Video Keyframe Retrieval	Michal Shlapentokh-Rothman, Prachi Garg, Yu-Xiong Wang	http://arxiv.org/abs/2605.23826v1	cs.CV, cs.CL	May 22, 2026	Detected	RDR84	Keyframe selection is a direct way to provide verifiable visual evidence for long-video quest...
Gated DeltaNet-2: Decoupling Erase and Write in Linear Attention	Ali Hatamizadeh, Yejin Choi, Jan Kautz	http://arxiv.org/abs/2605.22791v1	cs.AI	May 21, 2026	Detected	RDR84	Linear attention replaces the unbounded cache of softmax attention with a fixed-size recurren...
LCGuard: Latent Communication Guard for Safe KV Sharing in Multi-Agent Systems	Sadia Asif, Mohammad Mohammadi Amiri, Momin Abbas	http://arxiv.org/abs/2605.22786v1	cs.AI, cs.ET	May 21, 2026	None	RDR84	Large language model (LLM)-based multi-agent systems increasingly rely on intermediate commun...