A leave-one-out attention preprocessing scheme makes in-context learning permutation invariant without sacrificing access to other context examples, improving length and out-of-distribution generalization.
What learning algorithm is in-context learning? investigations with linear models
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Rethinking Invariance in In-context Learning
A leave-one-out attention preprocessing scheme makes in-context learning permutation invariant without sacrificing access to other context examples, improving length and out-of-distribution generalization.