"Representation geometry is a window into the inner world of neural networks."

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

• The geometry of internal representations and the geometry of model behavior share a precise correspondence — representation geometry is a window into the inner world of neural networks.claim0.907
  The paper's deepest interpretive claim, asserting that representation structure and behavioral structure are not coincidentally aligned but deeply connected.
• Geometry in neural representation is not merely incidental, but is in fact the proper object for enabling principled control via intervention on internals.claim0.862
  Core interpretive assertion: geometric structure is causally load-bearing, not epiphenomenal.
• Neural representation geometry causally shapes behavior; interventions respecting that geometry will yield natural trajectories.hypothesis0.846
  Central hypothesis tested via manifold steering experiments across language models and video world models.
• We hypothesize that representation geometry drives model behavior — the geometric structure of internal representations causally shapes what models do externally.hypothesis0.827
  The causal hypothesis motivating the use of causality (intervention) as the lens connecting representation and behavior geometry.
• Linear representation hypothesis: neural networks represent meaningful concepts as directions in their activation spaces.hypothesis0.825
  Foundation for interpreting features as linear directions.
• There exists a bidirectional relationship between the geometry of neural representation and the geometry of model behaviorclaim0.824
  Central empirical claim of the paper, demonstrated across tasks and modalities
• We hypothesize that a Representation Network (RN) emerges from LLM representations, where each dimension is a node and latent connections exist between nodes or clusters of nodes.hypothesis0.817
  Core methodological hypothesis enabling the application of IIT to LLM representation sequences.
• Neural Representation Geometryconcept0.816
  The broader conceptual framework that neural activations exhibit non-Euclidean geometric structure causally linked to behavior.