MarCognity-AI for MultiNRC
#2
by
elly99
- opened
Can moral reasoning be scaffolded cognitively?
MultiNRC explores moral judgment across cultures and contexts.
I’m building MarCognity-AI —https://huggingface.co/elly99/MarCognity-AI
a framework that scaffolds reflection before response in LLMs.
Could this be a complementary layer to rubric-based evaluation?
What if discomfort, ambiguity, and ethical tension became signals — not noise?
Would love to hear thoughts on integrating reflective checkpoints into moral reasoning pipelines.