LLMs trained on reconstructing known research ideation trajectories from citations outperform those inferring novel directions, yielding better novelty, diversity, and downstream research artifact quality.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Structure Liberates: How Constrained Sensemaking Produces More Novel Research Output
LLMs trained on reconstructing known research ideation trajectories from citations outperform those inferring novel directions, yielding better novelty, diversity, and downstream research artifact quality.