Modelpedia
Work in progress
About
Findings
Models
Concepts
Methods
Datasets
Sources
Light
Dark
Analyzing Feed-Forward Blocks in Transformers through the Lens of Attention Maps
2024-01-16
· ICLR 2024 spotlight ·
anchor
Findings
IC-1409
FF blocks in BERT and GPT-2 modify token-to-token contextualization, with the effect concentrated in specific layers and targeting specific linguistic compositions rather than simple word co-occurrence
IC-1410
FF's contextualization effects in BERT and GPT-2 are largely canceled by the residual connection and layer normalization, with LN's γ weights specifically shrinking the outlier dimensions in FF output