Latent Positional Information is in the Self-Attention Variance of Transformer Language Models Without Positional Embeddings
The use of positional embeddings in transformer language models is widely accepted. However, recent research has called into question the necessity of such embeddings. We further extend this inquiry by demonstrating that a randomly initialized and frozen transformer language model, devoid of positio...
Gespeichert in:
Hauptverfasser: | , , , , |
---|---|
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext bestellen |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | The use of positional embeddings in transformer language models is widely
accepted. However, recent research has called into question the necessity of
such embeddings. We further extend this inquiry by demonstrating that a
randomly initialized and frozen transformer language model, devoid of
positional embeddings, inherently encodes strong positional information through
the shrinkage of self-attention variance. To quantify this variance, we derive
the underlying distribution of each step within a transformer layer. Through
empirical validation using a fully pretrained model, we show that the variance
shrinkage effect still persists after extensive gradient updates. Our findings
serve to justify the decision to discard positional embeddings and thus
facilitate more efficient pretraining of transformer language models. |
---|---|
DOI: | 10.48550/arxiv.2305.13571 |