Following this puts new work involving it at the top of your briefing, with a note saying why it is there. Links are taken from the source record, never inferred.
A sound methodological study demonstrating improved prediction accuracy across multiple cancer cohorts using a novel filtering approach, but limited by lack of prospective validation, absence of hard clinical outcomes, and no comparison to established biomarkers.
A training-free algorithmic method for KV cache compression in reasoning models, demonstrated on benchmarks with reported memory and throughput improvements, but lacking peer review and clinical/clinical-adjacent validation.