Concept Boundary Vectors
This work addresses the need for better interpretability in machine learning models, particularly for researchers and practitioners seeking to analyze model representations, but it appears incremental as it builds on existing concept vector methods.
The authors tackled the problem of understanding and improving the interpretability of machine learning models by introducing concept boundary vectors, which are derived from the boundary between latent representations of concepts, and demonstrated that these vectors capture semantic meaning and are effective compared to concept activation vectors.
Machine learning models are trained with relatively simple objectives, such as next token prediction. However, on deployment, they appear to capture a more fundamental representation of their input data. It is of interest to understand the nature of these representations to help interpret the model's outputs and to identify ways to improve the salience of these representations. Concept vectors are constructions aimed at attributing concepts in the input data to directions, represented by vectors, in the model's latent space. In this work, we introduce concept boundary vectors as a concept vector construction derived from the boundary between the latent representations of concepts. Empirically we demonstrate that concept boundary vectors capture a concept's semantic meaning, and we compare their effectiveness against concept activation vectors.