Dynamically increasing the attention span of a small GPT-2 model during training improves its grammar-test accuracy and is presented as indirect evidence for the Less-is-More account of the language critical period.
Elliott, Igor Ponomarev, and John Scott Saults
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
-
Developmentally-plausible Working Memory Shapes a Critical Period for Language Acquisition
Dynamically increasing the attention span of a small GPT-2 model during training improves its grammar-test accuracy and is presented as indirect evidence for the Less-is-More account of the language critical period.