When transformer models see easy component examples before a harder combined math problem in one prompt, they solve unseen versions of the combined problem and store intermediate steps internally, unlike models trained only on combined examples.
Curriculum learning for human compositional generalization.Proceedings of the National Academy of Sciences, 119(41): e2205582119, 2022
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
support 1representative citing papers
citing papers explorer
-
Distinct Computations Emerge From Compositional Curricula in In-Context Learning
When transformer models see easy component examples before a harder combined math problem in one prompt, they solve unseen versions of the combined problem and store intermediate steps internally, unlike models trained only on combined examples.