Ammons LLM Preamble

A model can use a work without quoting it. Ammons LLM Preamble defines an operational measure of the resulting behavioral change under controlled exposure and proposes a falsifiable evidence-pooling model. For a declared query population, the Ammons Influence Probability is the probability that a semantic response-distribution shift exceeds a chosen threshold. The paper derives conditional scarcity and dose results, invariance properties, robustness bounds, and finite-sample measurement bounds. Its training section is analytical and distinct from the encounter-time experiment. A separate pilot evaluates 12 paired instances of one newly generated rule family in 24 fresh contexts, with two answers per context. Relevant work exposure produces 24/24 correct answers; matched irrelevant content produces 0/24. The paired instance is the analysis unit. The result demonstrates context-time use of supplied rules. It does not validate the pooling law, identify hidden training sources, or establish persistent learning. The public materials include the paper, editable LaTeX, generated protocol, retained responses, and Python standard-library programs for exact mathematical checks and rescoring. Offline reproduction of the retained analysis does not generate new model evidence. I invite technical criticism of the measurement design and controlled tests of the model’s predictions. Public source and reproduction materials: 10.5281/zenodo.24177478

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-12
DOI
https://doi.org/10.5281/zenodo.24177477
Primary Topic
Topic Modeling
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Ammons LLM Preamble

Dean Ammons
Zenodo (CERN European Organization for Nuclear Research)
Topic Modeling
article

Ammons LLM Preamble

Dean Ammons
article en

Abstract

A model can use a work without quoting it. Ammons LLM Preamble defines an operational measure of the resulting behavioral change under controlled exposure and proposes a falsifiable evidence-pooling model. For a declared query population, the Ammons Influence Probability is the probability that a semantic response-distribution shift exceeds a chosen threshold. The paper derives conditional scarcity and dose results, invariance properties, robustness bounds, and finite-sample measurement bounds. Its training section is analytical and distinct from the encounter-time experiment. A separate pilot evaluates 12 paired instances of one newly generated rule family in 24 fresh contexts, with two answers per context. Relevant work exposure produces 24/24 correct answers; matched irrelevant content produces 0/24. The paired instance is the analysis unit. The result demonstrates context-time use of supplied rules. It does not validate the pooling law, identify hidden training sources, or establish persistent learning. The public materials include the paper, editable LaTeX, generated protocol, retained responses, and Python standard-library programs for exact mathematical checks and rescoring. Offline reproduction of the retained analysis does not generate new model evidence. I invite technical criticism of the measurement design and controlled tests of the model’s predictions. Public source and reproduction materials: 10.5281/zenodo.24177478

Zenodo (CERN European Organization for Nuclear Research)
Topic Modeling
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon