Axiom Drift AI focuses on the structural dynamics of long-form human–LLM interaction.
The work asks a narrow question: what measurable structure emerges when humans and language models interact over hundreds of turns and what can that structure reveal about the interaction itself?
The approach is non-semantic. Measurements are derived from temporal organization, coupling behavior, and structural properties of the interaction without reference to content, sentiment, or meaning. Findings are published openly and instruments are publicly available.
| Title | Date | DOI | |
|---|---|---|---|
| 01 | Temporal Dynamics of Long-Form Human-LLM Interaction | Jan 2026 | 10.5281/zenodo.18273459 |
| 02 | Surface Matching, Temporal Divergence, and the Limits of Synthetic Baselines | Mar 2026 | 10.5281/zenodo.19143515 |
| 03 | Shared Temporal Structure and Distinct Coupling Regimes in Two Human-LLM Archives | Apr 2026 | 10.5281/zenodo.19633916 |
Variance Register — structural analysis of fixed example archives
Open Register — structural analysis of user-supplied transcripts · CSV · JSON · Markdown · plain text
TwinTrack — human archive vs. synthetic input regime comparison
All instruments run entirely in the browser. No data is transmitted or stored.