Dual Triangle Attention: Effective Bidirectional Attention Without Positional Embeddings

TL;DR

Dual Triangle Attention introduces a bidirectional attention mechanism that encodes positional information without explicit positional embeddings, enhancing sequence modeling across various domains.

Contribution

It proposes a novel bidirectional attention method using complementary triangular masks, eliminating the need for positional embeddings in transformers.

Findings

Dual Triangle Attention learns positional information without explicit embeddings.

Performs well in masked language modeling tasks on language and protein data.

Achieves the best context extension performance with Rotary Positional Embeddings.

Abstract

Bidirectional transformers are the foundation of many sequence modeling tasks across natural, biological, and chemical language domains, but they are permutation-invariant without explicit positional embeddings. In contrast, unidirectional attention inherently encodes positional information through its triangular mask, enabling models to operate without positional embeddings altogether. Here, we introduce Dual Triangle Attention, a novel bidirectional attention mechanism that separates the query-key subspace of each attention head into two complementary triangular masks: one that attends to past-and-self positions and one that attends to future-and-self positions. This design provides bidirectional context while maintaining the causal mask's implicit positional inductive bias in both directions. Using PyTorch's flex_attention, Dual Triangle Attention is implemented as a single compiled kernel call with no additional parameters beyond standard multi-head attention. We evaluated Dual Triangle Attention across three settings: (1) a synthetic argmax position probe, (2) masked language modeling (MLM) on natural language, and (3) MLM on protein sequences. In the argmax task, both Dual Triangle Attention and causal attention learn positional information without explicit positional embeddings, whereas standard bidirectional attention cannot. In the MLM experiments, Dual Triangle Attention with Rotary Positional Embeddings (RoPE) achieved the best context extension performance and strong performance across the board. These findings suggest that Dual Triangle Attention is a viable attention mechanism for bidirectional transformers, with or without positional embeddings.