Linear Direction, Conceptor (soft ellipsoidal region)

Functional Subspace, where language models can use vector algebra to solve problems (2026), Steering Large Language Models using Conceptors: Improving Addition-Based Activation Engineering (2024), Function Vectors in Large Language Models (2024), In-Context Learning Creates Task Vectors (2023), Unifying Attention Heads and Task Vectors via Hidden State Geometry in In-Context Learning (2025), Do Different Prompting Methods Yield a Common Task Representation in Language Models? (2025), Mass-Editing Memory in a Transformer (2022), Eliciting Latent Predictions from Transformers with the Tuned Lens (2023)