Cognitive Fatigue in Autoregressive Transformers: Formalization and Measurement

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Marwah, Riju, Garimella, Ritvik, Pallagani, Vishal, Jain, Atishay, Stewart, Michael, Sheth, Amit
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866914616127258624
author Marwah, Riju
Garimella, Ritvik
Pallagani, Vishal
Jain, Atishay
Stewart, Michael
Sheth, Amit
author_facet Marwah, Riju
Garimella, Ritvik
Pallagani, Vishal
Jain, Atishay
Stewart, Michael
Sheth, Amit
contents Autoregressive language models frequently degrade during long-horizon generation, producing repetitive text, losing instruction adherence, and exhibiting unstable entropy. Despite the prevalence of these failures, practitioners lack online diagnostics to detect them in real-time as they occur. We formalize this degradation as cognitive fatigue, a measurable generation-time state characterized by decay in attention to the original prompt, representational drift, and entropy miscalibration. We introduce the Fatigue Index (FI), a lightweight, model-agnostic diagnostic that aggregates these three signals under explicit axioms (monotonicity, boundedness, interpretability) enabling reliable runtime monitoring. Across nine models (1B-13B parameters), FI trajectories exhibit structured temporal dynamics, predict task degradation (AUROC = 0.95) and repetition (Spearman rho = 0.94), and reveal non-monotonic scaling behavior: instruction-tuned models below 3B exhibit faster collapse than base models, with this trend reversing at 7B. Stress analyses further show that FI onset accelerates under longer contexts, middle-positioned evidence, and reduced numerical precision. These results establish cognitive fatigue as a coherent and measurable phenomenon, and position FI as a principled tool for runtime reliability monitoring in production LLM systems.
format Preprint
id arxiv_https___arxiv_org_abs_2605_30981
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Cognitive Fatigue in Autoregressive Transformers: Formalization and Measurement
Marwah, Riju
Garimella, Ritvik
Pallagani, Vishal
Jain, Atishay
Stewart, Michael
Sheth, Amit
Computation and Language
Machine Learning
Autoregressive language models frequently degrade during long-horizon generation, producing repetitive text, losing instruction adherence, and exhibiting unstable entropy. Despite the prevalence of these failures, practitioners lack online diagnostics to detect them in real-time as they occur. We formalize this degradation as cognitive fatigue, a measurable generation-time state characterized by decay in attention to the original prompt, representational drift, and entropy miscalibration. We introduce the Fatigue Index (FI), a lightweight, model-agnostic diagnostic that aggregates these three signals under explicit axioms (monotonicity, boundedness, interpretability) enabling reliable runtime monitoring. Across nine models (1B-13B parameters), FI trajectories exhibit structured temporal dynamics, predict task degradation (AUROC = 0.95) and repetition (Spearman rho = 0.94), and reveal non-monotonic scaling behavior: instruction-tuned models below 3B exhibit faster collapse than base models, with this trend reversing at 7B. Stress analyses further show that FI onset accelerates under longer contexts, middle-positioned evidence, and reduced numerical precision. These results establish cognitive fatigue as a coherent and measurable phenomenon, and position FI as a principled tool for runtime reliability monitoring in production LLM systems.
title Cognitive Fatigue in Autoregressive Transformers: Formalization and Measurement
topic Computation and Language
Machine Learning
url https://arxiv.org/abs/2605.30981