Metraipolis › archive › dc_2b29de1de08b
ActMap: Single-Pass Uncertainty Quantification from Generation-Time Activation Maps — arXiv:2609.11498
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.11498
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-16T05:24:24.440Z
- SHA-256 of the archived text
c211230e685718ba5cad6c5eac90f77c47ce8b5a02f046de4f90591ff576199d- Stable URL of this page
https://metraipolis.com/documento/dc_2b29de1de08b
How to cite
ActMap: Single-Pass Uncertainty Quantification from Generation-Time Activation Maps — arXiv:2609.11498 — arXiv (en), https://arxiv.org/abs/2609.11498 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_2b29de1de08b (Metraipolis archive dc_2b29de1de08b, sha256 c211230e685718ba5cad6c5eac90f77c47ce8b5a02f046de4f90591ff576199d)
BibTeX
@misc{dc2b29de1de08b,
title = {ActMap: Single-Pass Uncertainty Quantification from Generation-Time Activation Maps — arXiv:2609.11498},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.11498},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_2b29de1de08b; sha256 c211230e685718ba5cad6c5eac90f77c47ce8b5a02f046de4f90591ff576199d},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.11498. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
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 coordinates. The map is captured during the generation pass with no measurable overhead, has a fixed shape across model depths and hidden sizes, and occupies 96 KiB: a compact artifact that can be retained for audit-relevant generations and probed directly, with occlusion analysis localizing the classifier's signal to mid-depth regions of the map. A lightweight classifier, instantiated as a compact Vision Transformer, reads an estimated correctness probability from each map in a fraction of a millisecond; capacity-matched MLPs perform comparably, indicating the representation itself carries the result. Trained and evaluated in-domain on short-answer QA, direct-answer math, and summarization factuality with three instruction-tuned 7-8B models, ActMap consistently outperforms sampling, token-probability, attention, and embedding baselines, and matches ACT-ViT, a detector trained on dense activation tensors $67\times$ larger, at essentially the same mean AUROC with lower calibration error on ten of twelve pairs. The resulting score supports abstention, routing, and selective verification from a single generation, making it a practical primitive for scalable oversight of deployed models. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-11 in cs.AI). Categorie arXiv: Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.11498 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-16T05:17:52.029Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-11?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).