trust-first

Tag

Cards List
#trust-first

AXIOM: A Trust-First Neuro-Symbolic Execution Architecture for Verifiable Mathematical Reasoning

arXiv cs.AI · 2026-06-02 Cached

AXIOM is a trust-first neuro-symbolic execution architecture for mathematical reasoning where the LLM acts as a canonicalizer, rewriting natural language problems into schemas processed by a deterministic CAS pipeline, achieving 94.36% correctness with 100% trust on parseable queries.

0 favorites 0 likes
← Back to home

Submit Feedback