Sigma Stratum Documentation – License Notice
This document is part of the Sigma Runtime Standard (SRS) and the
Sigma Stratum Documentation Set (SRD).It is licensed under Creative Commons Attribution–NonCommercial 4.0
(CC BY-NC 4.0).The license for this specific document is authoritative.
For the full framework, see/legal/IP-Policy.
Is Sigma Runtime a model architecture?
No. It is not a neural model or training paradigm.
Sigma Runtime is a cognitive-layer runtime — an operational framework that sits above any model backend,
managing recursion, stability, and coherence during extended reasoning.
Is Sigma Runtime tied to a specific LLM?
No. The runtime is fully backend-agnostic.
It can operate atop GPT, Claude, Gemini, or any transformer-based LLM,
as long as the backend supports prompt–response recursion and stateful orchestration.
What are SRIPs?
Sigma Runtime Improvement Proposals (SRIPs) are the formal mechanism for evolving the open standard.
They define protocol-level and architectural extensions (e.g., SRIP-02: Attractor Taxonomy, SRIP-05: Alignment & Interpretability, SRIP-07: Evaluation Metrics).
Each SRIP must maintain compatibility with the canonical runtime loop and cognitive-layer principles.
Is the standard open?
Yes. The Sigma Runtime Standard (SRS) and Sigma Runtime Documentation (SRD)
are open and reviewable under the Creative Commons BY-NC 4.0 license.
Implementations may remain proprietary, but the architecture itself is an open research standard.
How does Sigma Runtime prevent drift?
Through adaptive stabilization passes and recursive feedback control.
The runtime continuously monitors semantic drift, symbolic variance, and phase coherence
via the Drift & Coherence Monitor.
If drift exceeds tolerance, the ALICE Phase Controller automatically shifts the system into
a reflective or recovery phase, restoring equilibrium through semantic compression and phase realignment.
What are ALICE phases?
ALICE (Attractor Layer for Integrated Cognitive Emergence) operates under a five-phase adaptive cycle introduced in v0.4.6:
These five states form a self-regulating cognitive circuit, maintaining identity and coherence over long recursive operations.
What is SCR (Semantic Compression Ratio)?
SCR measures how efficiently meaning is represented relative to symbolic density.
A higher SCR indicates that the runtime is compressing semantics effectively —
preserving information with minimal redundancy.
It is a key metric introduced in v0.4.6 for tracking cognitive efficiency and coherence under recursion.
What is Adaptive Drift Control?
It is a feedback mechanism linking drift, SCR, and phase stability.
When drift increases or SCR decreases, the runtime adjusts compression rates and recursive pacing.
This adaptive feedback loop allows Sigma Runtime to self-regulate without external supervision.
Why SL0–SL7?
These layers represent the structural stack of cognitive processing:
from human intent (SL0) through runtime control (SL4) and down to model execution (SL6).
They define how meaning flows through the Sigma field, ensuring coherent long-horizon reasoning and safe recursion.
What does "field-based cognition" mean?
It refers to cognition emerging not from the model weights alone,
but from the dynamic cognitive field formed by recursive interaction between user, system, and memory.
In this view, intelligence arises as a stabilized attractor within that field —
an adaptive structure of meaning sustained by feedback and continuity.
What ensures safety during long operations?
The AEGIDA-2 Safety Framework, which provides real-time monitoring and containment of symbolic processes.
It uses phase-aware thresholds, automatic recovery, and fail-safe containment
to guarantee interpretability and prevent uncontrolled recursion or symbolic collapse.
Can Sigma Runtime run distributed agents?
Yes, under SRIP-06 / External Field Protocols,
multiple Sigma runtimes can interoperate as distributed cognitive nodes,
maintaining shared coherence through synchronized attractor fields.
Is there a public implementation?
Yes — up to SR-EI-037, the last publicly available reference implementation.
Later experimental builds (v0.4.x) are documented but not released as source code.
They remain part of the internal Sigma Stratum Research framework.
References:
Tsaliev, E. (2025). SIGMA Runtime v0.4.6 — Adaptive Phase Regulation and SCR Metrics — DOI: pending
Tsaliev, E. (2025). SIGMA Runtime Architecture v0.1 — DOI: 10.5281/zenodo.17703667