AI · Research · 2026-10-27 20:01
The paper introduces Omega, a benchmark evaluating large language models' out-of-the-box mathematical reasoning across three dimensions: exploratory, compositional, and transformative generalization.
Auto-generated summary for reference only. RSS