Bag of Coins: A Statistical Probe into Neural Confidence Structures

TL;DR

The paper introduces the Bag-of-Coins (BoC) probe, a diagnostic tool for analyzing neural network confidence and uncertainty geometry, revealing architecture-dependent differences in in-distribution and out-of-distribution detection.

Contribution

The BoC probe provides a new non-parametric method for diagnosing neural confidence structures and understanding how different architectures encode uncertainty.

Findings

ViT shows clear ID/OOD separation with BoC

ResNet and RoBERTa exhibit overlapping uncertainty geometry

BoC improves calibration only for poorly calibrated models

Abstract

Modern neural networks often produce miscalibrated confidence scores and struggle to detect out-of-distribution (OOD) inputs, while most existing methods post-process outputs without testing internal consistency. We introduce the Bag-of-Coins (BoC) probe, a non-parametric diagnostic of logit coherence that compares softmax confidence $\hat p$ to an aggregate of pairwise Luce-style dominance probabilities $\bar q$, yielding a deterministic coherence score and a p-value-based structural score. Across ViT, ResNet, and RoBERTa with ID/OOD test sets, the coherence gap $\Delta=\bar q-\hat p$ reveals clear ID/OOD separation for ViT (ID ${\sim}0.1$-$0.2$, OOD ${\sim}0.5$-$0.6$) but substantial overlap for ResNet and RoBERTa (both ${\sim}0$), indicating architecture-dependent uncertainty geometry. As a practical method, BoC improves calibration only when the base model is poorly calibrated (ViT: ECE $0.024$ vs.\ $0.180$) and underperforms standard calibrators (ECE ${\sim}0.005$), while for OOD detection it fails across architectures (AUROC $0.020$-$0.253$) compared to standard scores ($0.75$-$0.99$). We position BoC as a research diagnostic for interrogating how architectures encode uncertainty in logit geometry rather than a production calibration or OOD detection method.