LaTIM: Measuring Latent Token-to-Token Interactions in Mamba Models
This provides interpretability for Mamba models, addressing a gap for researchers and practitioners using efficient long-context sequence modeling.
The authors tackled the lack of interpretability tools for state space models like Mamba by introducing LaTIM, a token-level decomposition method, which effectively reveals token-to-token interaction patterns across tasks such as machine translation and retrieval-based generation.
State space models (SSMs), such as Mamba, have emerged as an efficient alternative to transformers for long-context sequence modeling. However, despite their growing adoption, SSMs lack the interpretability tools that have been crucial for understanding and improving attention-based architectures. While recent efforts provide insights into Mamba's internal mechanisms, they do not explicitly decompose token-wise contributions, leaving gaps in understanding how Mamba selectively processes sequences across layers. In this work, we introduce LaTIM, a novel token-level decomposition method for both Mamba-1 and Mamba-2 that enables fine-grained interpretability. We extensively evaluate our method across diverse tasks, including machine translation, copying, and retrieval-based generation, demonstrating its effectiveness in revealing Mamba's token-to-token interaction patterns.