Show HN: LLM Crisis Triage Audit – Deterministic state-space governance (github.com)

A new repository has been launched featuring the empirical evaluation transcripts and zero-shot prompt logs for the research paper titled "Beyond Algorithmic Deflection: A Deterministic State-Space Triage Architecture for Crisis Intervention and Cognitive Ergonomics in Large Language Models." Authored by M.M.M. from Recursive-Logic-Core, the study investigates how large language models (LLMs) can be effectively governed during crises, contributing significantly to the intersection of AI/ML and cognitive ergonomics. This initiative is crucial for the AI/ML community as it aims to address the challenges posed by LLMs in crisis situations, ensuring they provide reliable and responsible outputs. The repository includes detailed audit logs for major AI models such as Anthropic Claude, Google Gemini, OpenAI's GPT, and xAI Grok, all of which are openly accessible under a Creative Commons license. This transparency is intended to promote replication studies and enhance mechanistic alignment, allowing researchers to build on this work for better governance frameworks in AI applications.