Four Layers of AI Understanding - Testable Instrument: Annotation Rubric, Pivot Rules, Case Bank (20), and Scoring Sheet (v0.1)
Year: 2025
Type: article
Abstract: This record publishes v0.1 of the Four-Layer (L1–L4) annotation rubric and its operational mapping to pivots under the L0-STEP framework (Station → Event → Pivot), including a 20-case template bank and a one-page scoring sheet. It is designed as a repeatable labelling instrument for structured human–LLM annotation and for reproducible event logging aligned with the LFP-0 (6+H) decision-record schema. This document is a specification and operational instrument, not a validated metric. It is conceptually distinct from the formally designed controlled DER reconstructability evaluation study proposed as the primary doctoral empirical unit. Any performance, reliability, or governance claims require separate empirical validation. Author: Hon Bor So ORCID: https://orcid.org/0009-0008-2768-7494 The author also publishes under the name Sing So (Chöndrel Dorje). Project entry point (Root Index): So, Hon Bor. (2025). Master Roadmap and Index for TPT & LFP-0 (Root Index v1.0). Zenodo. https://doi.org/10.5281/zenodo.17992916 License: CC BY-NC 4.0
Author: So, Hon Bor
Language: English
Topic: Topic Modeling
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7116341877