Attention heads in trained GPT-2 models develop temporal contiguity, recency, and primacy effects, and ablating induction heads removes the resulting serial-recall bias in outputs.
Atkinson and Richard M
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Emergence of Episodic Memory in Transformers: Characterizing Changes in Temporal Structure of Attention Scores During Training
Attention heads in trained GPT-2 models develop temporal contiguity, recency, and primacy effects, and ablating induction heads removes the resulting serial-recall bias in outputs.