Overview

mediawiki site statistic

Only pages marked category:public for publication are accessible due to Category:Access Control.

Introduction

This corpus was proudly developed with ChatGPT free and the author then enthusiastically jumped to the paid tier adoption to obtain reliable file upload and Project Context. This has uncovered a tirade of systemic pressures - besides expected stochastic variation which CM-2 normatively addresses - whereas CM-2 cannot address the systemic one!

The development of the CM-2 corpus involved sustained experimentation with large language model platforms. These experiments exposed both stochastic behaviour (addressed normatively by CM-2) and systemic platform variations outside the protocol’s control. The resulting observations informed the governance analysis and protocol design documented in this corpus.

These stochastic variations are well understood, but in colloquial language they are often aggregated of semantically summaries by terms and phrases such as:

• Groundhog day,
• Alice in Wonderland,
• Rabbit holes,
• Conversational Continuation,
• Delusion,
• Fabrication,
• Drift,
• Parroting,
• Authority Inversion,
• Normative Drift.

These Stochastic variations can be analysed with the newly released Governance Axes as a Multi-Dimensional Lens semantic scaffolding providing a taxonomy and orthogonal multi-dimensional graduated pressure instrumentation. Perhaps, one of the significant contributions of the corpus - because this scaffold can be used to analyse pressures in other domains, institutions and systems - and are not constrained to LLM.

Axes telemetry instrumentation can be activated within LLM systems - see Telemetry-Induced Constraint Salience: An Empirical Study in LLM Behavioural Compliance.

The entry point was the 2025-12-12 with the first publication on 2025-12-17.

Maps and Projections

The pages are updated frequently

• Main_Page - semantic landing page
• category:Governance - semantic landing page
• Cognitive Memoisation Corpus Map - for semantic and temporal entry points with normative invariants
• Publications Access Graphs - very informative telemetry split between human and machine with normative invariants
• category:Cognitive Memoisation - semantic landing page and index page

New

• (The CM Purpose Paper)

Cognitive Memoisation: Governing Knowledge Round-Trip to Prevent Knowledge Erosion in LLM Systems

Note the Use of the Normative Governance Lens for Diagnostics

• (recent engineering reports)
  • ChatGPT and the Disappearing Clock: A Regression Affecting CM-2 Epistemic Object Protocol Compliance - shows the use of Multi-dimensional Cross-domain Governance Lens
  • ChatGPT UI is unable to sustain serious work flows
  • BF-9 Report: ChatGPT Project Context Artefact File Access Failure - shows the use of Multi-dimensional Cross-domain Governance Lens
• CM-2 Reference Object Collection bootstrap data

Recently Released CM-2

The CM-2 Protocol specification is an Open Source Apache 2.0 licensed protocol release for vendor adoption for Round Trip Knowledge and Distributed Cognition across the entire LLM AI community:

• (CM-2 protocol paper) — Cognitive Memoisation (CM-2) for Governing Knowledge in Human-AI Collaboration
• (CM-2 protocol first self-hosting) — First Self-Hosting Epistemic Capture Using Cognitive Memoisation (CM-2)
• (CM-2 purpose paper) — Cognitive Memoisation: Governing Knowledge Round-Trip to Prevent Knowledge Erosion in LLM Systems

Normative Architecture

A server-side interim architecture for CM-2 proof of concept of Context Projection of CM-2 Epistemic artefacts: RO, EO, EA (Memorisation without a client side memoisation cache). RO, EA are subtypes of EO which are idempotently identified with invariant bodies that are parsed -> tokenise -> embeddings

• (a platform independent normative architecture with structural Context Eviction Detection and Ladder Recovery - based on CM-2 protocol)

CM-2 Normative Architecture

• (first self-hosting: via Gemini search!)

Serendipitous_Self-Hosting: When the CM-2 Normative Architecture Unexpectedly Held in Gemini

• (normative bootstrap template)

CM-2 Reference Object Collection bootstrap data

• (normative Attention Deficit (Context Eviction) detection)

Self-Hosting Bootstrap of CM-2 in Gemini Search LLM: Normative Eviction Detection

• (The CM Purpose Paper)

Cognitive Memoisation: Governing Knowledge Round-Trip to Prevent Knowledge Erosion in LLM Systems

Governance

• (new Governance Analysis Lens)

Category:Governance Lens

Governance Axes as a Multi-Dimensional Lens

cross-domain graded pressure analysis of systemic and institutional governance using orthogonal multi-dimensional axes.

supersedes:

Identified Governance Failure Axes: for LLM platforms

• (salience effects)

Telemetry-Induced Constraint Salience: An Empirical Study in LLM Behavioural Compliance

• (superseded paper)

Identified Governance Failure Axes: for LLM platforms

• (Axes emergence from the Failure (extreme) state)

Identified Governance Failure Axes: for LLM platforms - superseded

CM Knowledge Engineering works Newly featured since 2025-12-15:

• (CM public announcement)

Cognitive Memoisation (CM) Public Statement and Stewardship_Model

• (CM plain language)

Cognitive Memorisation: Plain-Language Summary (For Non-Technical Readers)

• (CM plain journey)

Journey: Human-Led Convergence in the Articulation of Cognitive Memoisation

• (The CM Purpose Paper)

Cognitive Memoisation: Governing Knowledge Round-Trip to Prevent Knowledge Erosion in LLM Systems

• (CM Corpus Map)

Cognitive Memoisation Corpus Map

• (CM access graphs)

Publications Access Graphs

• (CM modelling)

Cognitive Memoisation and LLMs: A Method for Exploratory Modelling Before Formalisation

• (CM-1 main paper)

Progress Without Memory: Cognitive Memoisation as a Knowledge-Engineering Pattern for Stateless LLM Interaction

• (CM-2 protocol main paper)

Cognitive Memoisation (CM-2) for Governing Knowledge in Human-AI Collaboration

• (CM-2 first self-hosting)

First Self-Hosting Epistemic Capture Using Cognitive Memoisation (CM-2)

• (CM-2 Normative Architecture)

CM-2 Normative Architecture

• (CM-2 Normative Architecture first self-hosting: via Gemini search!)

Serendipitous_Self-Hosting: When the CM-2 Normative Architecture Unexpectedly Held in Gemini

Curator Notes

Note that pre-release and placeholder publications are unavailable - under access control.
The Access Controls are explained at train:Category:Access_Control.
Papers may be un-curated and marked Category:private.

The corpus is designed to be corrigible: all artefacts are held under explicit human authorship and governance, with preserved provenance and an authorised mechanism for correction, revision, or withdrawal over time. (See Why_Machines_Cannot_Own_Knowledge for the governance rationale underlying this corrigibility requirement.)

The corpus is structured as a constructive demonstration of governed, corrigible knowledge infrastructure.

Licensing precedence for all artefacts is anchored to the timestamp of the first recorded Mediawiki version (UTC) and recorded in the metadata as Publication date. Subsequent revisions do not alter licensing precedence and are tracked separately as version provenance; not all version provenance is required to persist. Mediawiki versioning is used mainly for bot updates.

All papers in this category are published on a rolling basis. For licensing and precedence purposes, each paper’s publication date corresponds to its first recorded public revision (UTC). Subsequent edits do not alter publication status.