Displacement Is Not Direction: Evaluating Fidelity Metrics for Quantized LLM Deployment

Abstract

Fidelity metrics, such as per-token KL divergence (KLD) against a high-precision reference, are often used in practice as low-cost proxies for benchmark quality. We test this practice on a 28-quant cohort of Qwen3.6-35B-A3B and a 41-quant cohort of Devstral-Small-2-24B, evaluated across a suite of downstream benchmarks. We find that KLD is strongly correlated with benchmark score over the full cohort ($ρ=-0.72$ on Qwen and $ρ=-0.86$ on Devstral, both with $p<0.001$). However, this relationship collapses to non-significance in the near-baseline silent zone ($ρ=+0.00$ on Qwen and $ρ=-0.24$, $p=0.36$, on Devstral). This collapse persists across 14 measurement variants, including different KLD aggregations, perplexity formulations, top-1 agreement, calibration corpora, and context lengths. At the per-prompt level, KLD has only weak failure-prediction power on code, with failed-vs-passed geometric-mean ratios in $[1.08,1.22]$ across five models on LiveCodeBench, and fails as a cross-model router, achieving only $42.3\%-49.4\%$ accuracy on disagreement prompts. We trace the collapse to a structural decomposition: KLD primarily measures the volume of disagreement with the reference, with silent-zone composite $ρ=+0.94$ ($p<0.001$) on Qwen and $+0.55$ ($p=0.03$) on Devstral, while its relationship to the direction of those disagreements is weak and task-conditional.