Mechanistic Interpretability of Cognitive Complexity in LLMs via Linear Probing using Bloom's Taxonomy
This provides a novel evaluation framework for interpreting cognitive complexity in LLMs, which is incremental as it builds on existing probing methods.
The study tackled the problem of evaluating cognitive complexity in Large Language Models by using Bloom's Taxonomy as a hierarchical lens, and found that linear classifiers achieved approximately 95% mean accuracy across all cognitive levels, indicating that cognitive level is encoded in a linearly accessible subspace of the models' representations.
The black-box nature of Large Language Models necessitates novel evaluation frameworks that transcend surface-level performance metrics. This study investigates the internal neural representations of cognitive complexity using Bloom's Taxonomy as a hierarchical lens. By analyzing high-dimensional activation vectors from different LLMs, we probe whether different cognitive levels, ranging from basic recall (Remember) to abstract synthesis (Create), are linearly separable within the model's residual streams. Our results demonstrate that linear classifiers achieve approximately 95% mean accuracy across all Bloom levels, providing strong evidence that cognitive level is encoded in a linearly accessible subspace of the model's representations. These findings provide evidence that the model resolves the cognitive difficulty of a prompt early in the forward pass, with representations becoming increasingly separable across layers.