# Understanding Softmax Confidence and Uncertainty Source: [Understanding Softmax Confidence and Uncertainty]() Collection index: [llms.txt]() Data updated: 2025-10-25T23:16:37.291Z Authors: Tim Pearce, Alexandra Brintrup, Jun Zhu Published: 2021-06-09T10:37:29.000Z ## Abstract It is often remarked that neural networks fail to increase their uncertainty when predicting on data far from the training distribution. Yet naively using softmax confidence as a proxy for uncertainty achieves modest success in tasks exclusively testing for this, e.g., out-of-distribution (OOD) detection. This paper investigates this contradiction, identifying two implicit biases that do encourage softmax confidence to correlate with epistemic uncertainty: 1) Approximately optimal decision boundary structure, and 2) Filtering effects of deep networks. It describes why low-dimensional intuitions about softmax confidence are misleading. Diagnostic experiments quantify reasons softmax confidence can fail, finding that extrapolations are less to blame than overlap between training and OOD data in final-layer representations. Pre-trained/fine-tuned networks reduce this overlap. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Liver Fibrosis Staging on Cirrhosis S4 vs S1-3](): Liver fibrosis staging; Cirrhosis S4 vs S1-3; Accuracy: 77.1. - [Liver fibrosis staging on Substantial Fibrosis S1 vs S2-4](): Liver fibrosis staging; Substantial Fibrosis S1 vs S2-4; Accuracy: 78.2. ## Original sources and implementation [Original paper]() [PDF]()