Mysticbirdie/hallucination-elimination-benchmark

Multi-tier benchmark: Cultural grounding + Triad Engine eliminates LLM hallucination across Claude 4.6, GPT-5.2, Mistral 7B, Gemini 2.5 Pro. Raw 15-58% → 95-100% accuracy on 222 adversarial QA pairs (Ancient Rome 110 CE). Novel topological paradox detection (F1=0.939, zero-shot). Model-agnostic, in production.

2 best points · 3 comments · 2 Hacker News posts · 7 GitHub stars · View on GitHub

Hacker News discussions

• Show HN: 100% LLM accuracy–no fine-tuning, JSON only (2 points, 2 comments, by MysticBirdie, 2026-02-25)
• Claude Code Mexico breach: training safety failed ground truth layer (2 points, 1 comments, by MysticBirdie, 2026-02-26)