Paraphrases of an identity document induce tighter clustering in LLM activation space than matched controls, indicating attractor-like dynamics for agent identity.
Transformerdynamics: Aneuroscientificapproachtointerpretabil- ity of large language models.arXiv preprint arXiv:2502.12131
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2representative citing papers
Training installs a depth-dependent spectral gradient and low-rank bottleneck in LLM residual streams whose amplification or suppression of graph communities is predicted by local operator type.
citing papers explorer
-
Identity as Attractor: Geometric Evidence for Persistent Agent Architecture in LLM Activation Space
Paraphrases of an identity document induce tighter clustering in LLM activation space than matched controls, indicating attractor-like dynamics for agent identity.
-
Dynamics of the Transformer Residual Stream: Coupling Spectral Geometry to Network Topology
Training installs a depth-dependent spectral gradient and low-rank bottleneck in LLM residual streams whose amplification or suppression of graph communities is predicted by local operator type.