elly99-AI/MarCognity-AI
View on GitHub ↗ ★ 8 Jupyter Notebook
A research framework for structured LLM evaluation, claim verification and reflective reasoning mechanisms.
4 mentions on Hacker News by 1 person: 4 stories, 0 comments · first 2026-03-17 · latest 2026-04-29
20082026
-
▲ 1
MarCognity-AI – epistemic layer for LLMs, runs locally, no APIstory by elly-99 · 2026-04-29 · 0 comments
-
▲ 1
MarCognity-AI v2.7.0 – factual grounding for LLM epistemic verificationstory by elly-99 · 2026-04-02 · 0 comments
-
▲ 2
MarCognity-AI – Epistemic failure rates in LLMs across 8 domainsstory by elly-99 · 2026-03-29 · 0 comments
-
▲ 3
A structural epistemic limit in LLMs: 8–15% unverifiable claims across domainsstory by elly-99 · 2026-03-17 · 0 comments