SOURCE-LINKED INTELLIGENCE

Perplexity Cost Understates What Activation Quantisation Breaks

arXiv · AI, language, vision and robotics · article · Sep 19, 2026 · UTC

Activation quantisation is usually evaluated with an aggregate metric, perplexity, averaged over every token a model predicts. We ask whether that average identifies which computations a quantiser damages. Perplexity turns out to be a reliable aggregate signal: across 12 models from four families and 780 within-model comparisons, the arm perplexity prefers also retains more induction and more retrieval in all but 2.1 and 4.0 percent of cases respectively. But where perplexity has risen by only a factor of 1.2 to 1.5, induction still keeps 0.959 of its intact accuracy while retrieval has alread

Read original source ↗ Open in workspace

recordType

paper

region

Global