LSC and Massively Documented LLM Hallucination: Prompt Archive, Simulation Lab, and Model-Lineage Extension

This continuation extends the MDLH / MHLM archive with a public prompt archive, reproducible simulation layer, and model-lineage auditing materials. It adds the Prompt Archive for preserving model instructions and iteration design, the LLM Simulation Lab for testing apparent coherence versus validation pressure, a static Model Lineage Simulator, and a report summarizing how GPT, Manus, Gemini, Codex, DeepSeek, and the LSC 6.2.2 correction layer shaped, amplified, constrained, or repaired the LSC theory line. The update does not claim that the underlying LSC physics is validated. It preserves the dual interpretation of the original record: LSC may be treated as an unvalidated physics hypothesis requiring independent falsification, and the same archive may also serve as a documented AI-safety case study in Massively Documented LLM Hallucination. The prompt and simulation layers track model instructions, apparent coherence, validation pressure, disagreement distance, and hallucination-risk pressure. These values are working variables for audit and reproducibility, not empirical proof.