Unstructured natural-language reasoning plans outperform dynamically generated JSON reasoning plans in an instance-level Self-Discover framework, with relative gains up to 18.90% on MATH.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Effects of structure on reasoning in instance-level Self-Discover
Unstructured natural-language reasoning plans outperform dynamically generated JSON reasoning plans in an instance-level Self-Discover framework, with relative gains up to 18.90% on MATH.