Director-Class AI: Coherence Engine — AI Output Verification & Safety Oversight
10.5281/zenodo.19244060 , 10.5281/zenodo.19118149 , 10.5281/zenodo.19210775 , 10.5281/zenodo.18842249 , 10.5281/zenodo.18896162 , 10.5281/zenodo.19102723 , 10.5281/zenodo.19224599 , 10.5281/zenodo.19267954 , 10.5281/zenodo.18854703 , 10.5281/zenodo.19430245 , 10.5281/zenodo.19140722 , 10.5281/zenodo.19091926 , 10.5281/zenodo.18840841 , 10.5281/zenodo.18822226 , 10.5281/zenodo.18897194 , 10.5281/zenodo.18837753 , 10.5281/zenodo.18828336 , 10.5281/zenodo.18843554 , 10.5281/zenodo.18822155 , 10.5281/zenodo.18928898 , 10.5281/zenodo.19210847 , 10.5281/zenodo.18828836 , 10.5281/zenodo.18835580 , 10.5281/zenodo.18928897 , 10.5281/zenodo.18822156 , 10.5281/zenodo.18896151 , 10.5281/zenodo.18897798 , 10.5281/zenodo.18841133 , 10.5281/zenodo.19172428
10.5281/zenodo.19244060 , 10.5281/zenodo.19118149 , 10.5281/zenodo.19210775 , 10.5281/zenodo.18842249 , 10.5281/zenodo.18896162 , 10.5281/zenodo.19102723 , 10.5281/zenodo.19224599 , 10.5281/zenodo.19267954 , 10.5281/zenodo.18854703 , 10.5281/zenodo.19430245 , 10.5281/zenodo.19140722 , 10.5281/zenodo.19091926 , 10.5281/zenodo.18840841 , 10.5281/zenodo.18822226 , 10.5281/zenodo.18897194 , 10.5281/zenodo.18837753 , 10.5281/zenodo.18828336 , 10.5281/zenodo.18843554 , 10.5281/zenodo.18822155 , 10.5281/zenodo.18928898 , 10.5281/zenodo.19210847 , 10.5281/zenodo.18828836 , 10.5281/zenodo.18835580 , 10.5281/zenodo.18928897 , 10.5281/zenodo.18822156 , 10.5281/zenodo.18896151 , 10.5281/zenodo.18897798 , 10.5281/zenodo.18841133 , 10.5281/zenodo.19172428
Director-Class AI: Coherence Engine — AI Output Verification & Safety Oversight
Real-time LLM hallucination guardrail. Dual-entropy scoring (NLI + RAG) with token-level streaming halt. 75.86% BA on AggreFact. Meta-confidence scoring, cross-turn contradiction tracking, structured output verification (JSON, tool calls, code), and online calibration from production feedback. Zero new dependencies for verification (stdlib only).
Real-time hallucination guardrail by ANULUM. Token-level streaming oversight with NLI + RAG fact-checking.
AI safety, streaming halt, DeBERTa, coherence engine, NLI verification, FactCG, hallucination detection, LLM guardrail, RAG fact-checking
AI safety, streaming halt, DeBERTa, coherence engine, NLI verification, FactCG, hallucination detection, LLM guardrail, RAG fact-checking
selected citations These citations are derived from selected sources.This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).0 popularity This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.Average influence This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).Average impulse This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.Average
Citations provided by BIP!Popularity provided by BIP!