SILENCIUM: A Pre-Inference De-Escalation and Intent-Gating Framework for LLM Interfaces (including Technical Addendum on Quantitative Drift Measurement)
Fuente:
Zenodo
Guardado en:
| Autor principal: | |
|---|---|
| Formato: | Recurso digital |
| Lenguaje: | inglés |
| Publicado: |
Zenodo
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866901648363749376 |
|---|---|
| author | Nowak, Daniel |
| author_facet | Nowak, Daniel |
| contents | <p class="p1"><span class="s1">Abstract:</span></p> <p class="p1"><span class="s2">This repository contains the complete </span><span class="s1">SILENCIUM</span><span class="s2"> framework, comprising the conceptual White Paper (Part I) and the Technical Addendum (Part II). </span><span class="s1">SILENCIUM</span><span class="s2"> is an innovative approach to Interaction Governance in Large Language Model (LLM) interfaces. It addresses the challenges of escalation dynamics and manipulative exploitation of AI systems by introducing a Pre-Inference De-Escalation and Intent-Gating framework.</span></p> <p class="p1"><span class="s1">Core Components of the Framework:</span></p> <p class="p2"><span class="s1">Intent-Gating:</span><span class="s2"> A mechanism to filter user queries based on the identification of rhetorical pressure and emotional manipulation, preventing unnecessary inference costs and boundary violations.</span></p> <p class="p2"><span class="s1">De-Escalation Policy:</span><span class="s2"> Strategies for the professional limitation of system responses when boundaries are crossed, inspired by proven de-escalation techniques from high-stress environments.</span></p> <p class="p2"><span class="s1">Quantitative Drift Measurement (Part II):</span><span class="s2"> Formal mathematical definitions and metrics for measuring </span><span class="s1">Semantic Drift</span><span class="s2"> and </span><span class="s1">Epistemic Drift</span><span class="s2">. It provides a methodology to quantify the divergence between user intent, system instructions, and factual grounding (Composite Risk Score).</span></p> <p class="p1"><span class="s2">This framework is designed for developers, safety researchers, and AI infrastructure providers seeking to implement more stable, cost-effective, and secure interaction designs.</span></p> |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_5281_zenodo_19812746 |
| institution | Zenodo |
| language | eng |
| publishDate | 2026 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | SILENCIUM: A Pre-Inference De-Escalation and Intent-Gating Framework for LLM Interfaces (including Technical Addendum on Quantitative Drift Measurement) Nowak, Daniel ai safety LLM governance Behavioral AI Intent-Gating Semantic Drift Epistemic Drift Interaction Design De-escalation Pre-Inference Filtering <p class="p1"><span class="s1">Abstract:</span></p> <p class="p1"><span class="s2">This repository contains the complete </span><span class="s1">SILENCIUM</span><span class="s2"> framework, comprising the conceptual White Paper (Part I) and the Technical Addendum (Part II). </span><span class="s1">SILENCIUM</span><span class="s2"> is an innovative approach to Interaction Governance in Large Language Model (LLM) interfaces. It addresses the challenges of escalation dynamics and manipulative exploitation of AI systems by introducing a Pre-Inference De-Escalation and Intent-Gating framework.</span></p> <p class="p1"><span class="s1">Core Components of the Framework:</span></p> <p class="p2"><span class="s1">Intent-Gating:</span><span class="s2"> A mechanism to filter user queries based on the identification of rhetorical pressure and emotional manipulation, preventing unnecessary inference costs and boundary violations.</span></p> <p class="p2"><span class="s1">De-Escalation Policy:</span><span class="s2"> Strategies for the professional limitation of system responses when boundaries are crossed, inspired by proven de-escalation techniques from high-stress environments.</span></p> <p class="p2"><span class="s1">Quantitative Drift Measurement (Part II):</span><span class="s2"> Formal mathematical definitions and metrics for measuring </span><span class="s1">Semantic Drift</span><span class="s2"> and </span><span class="s1">Epistemic Drift</span><span class="s2">. It provides a methodology to quantify the divergence between user intent, system instructions, and factual grounding (Composite Risk Score).</span></p> <p class="p1"><span class="s2">This framework is designed for developers, safety researchers, and AI infrastructure providers seeking to implement more stable, cost-effective, and secure interaction designs.</span></p> |
| title | SILENCIUM: A Pre-Inference De-Escalation and Intent-Gating Framework for LLM Interfaces (including Technical Addendum on Quantitative Drift Measurement) |
| topic | ai safety LLM governance Behavioral AI Intent-Gating Semantic Drift Epistemic Drift Interaction Design De-escalation Pre-Inference Filtering |
| url | https://doi.org/10.5281/zenodo.19812746 |