List your company

Mathematical Reasoning on BIG-Bench Hard Multi-Step Arithmetic Two (OOD)

12.4AccuracyLOVER
11.56811.7841212.216May 7, 2026

Evaluation results

MethodLinks
LOVERBackbone=mistral-7b2026.0512.4
Majority VotingBackbone=mistral-7b2026.0512
CoT-DecodingBackbone=mistral-7b2026.0511.6