SOURCE-LINKED INTELLIGENCE

ActMap: Single-Pass Uncertainty Quantification from Generation-Time Activation Maps

arXiv · AI, language, vision and robotics · article · Sep 10, 2026 · UTC

Practical uncertainty quantification (UQ) for large language models must decide, from a single generation, whether a specific answer should be trusted. Existing methods either sample multiple generations, read only output-token probabilities, or reduce the model's internal computation to a single hidden state. We introduce ActMap, a white-box representation that compresses the generation-time hidden-state trajectory (every layer, every generated token) into a fixed $12\times32\times128$ tensor of temporal-statistic channels that preserves structure across transformer depth and pooled hidden co

Read original source ↗ Open in workspace

recordType

paper

region

Global