MIRA-Math introduces a 2,310-instance benchmark isolating the ability of LLMs to request a single missing atomic fact needed to solve an underdetermined mathematical problem and then integrate it into an exact answer.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
MIRA-Math: A Benchmark for Minimal Information Requesting and Mathematical Reasoning
MIRA-Math introduces a 2,310-instance benchmark isolating the ability of LLMs to request a single missing atomic fact needed to solve an underdetermined mathematical problem and then integrate it into an exact answer.