This dossier bundles the Research desk's ongoing digest series: each issue, the editorial team selects three to five new arXiv preprints from the past 24 to 48 hours in cs.AI, cs.LG, and cs.CL that concretely touch the development or deployment of AI systems – from agent safety to training methods to evaluation gaps.
Selection follows traceable methodology in the abstract, solid numeric results, and thematic spread rather than repeated focus areas. All cited preprints are unreviewed at the time of their respective issue; how well their findings hold up is for independent replications to show.
