SAGE makes modeling strategy explicit during dataset construction and post-training with Segment-Weighted GRPO, raising average pass@1 from 72.7 to 80.3 while increasing formulation diversity and producing 14.2% more compact constraint systems.
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
UNVERDICTED 1representative citing papers
citing papers explorer
-
Strategy-Aware Optimization Modeling with Reasoning LLMs
SAGE makes modeling strategy explicit during dataset construction and post-training with Segment-Weighted GRPO, raising average pass@1 from 72.7 to 80.3 while increasing formulation diversity and producing 14.2% more compact constraint systems.