Top ten variables shown in descending order of rank. Variable categories were denoted by W (wetland), U (upland), P (precipitation), and T (temperature). Signs indicated the relationship between the p
New finding: The SHAPE of per-layer SAE feature coherence distributions discriminates hallucination better than any scalar summary (K, C, C_delta). Hallucination is not a depth shift — it is a coheren
How can we interpret predictions of a workload classification model? A workload is a sequence of operations executed in DRAM, where each operation contains a command and an address. Classifying a give
