Do Large Language Models Play Six Degrees of Separation? Measuring Topological Compression in Long-Context Manifolds

Excerpt

arXiv:2608.17950v1 Announce Type: new Abstract: Large Language Models (LLMs) demonstrate remarkable multi-hop reasoning capabilities over long contexts, yet the internal mechanisms enabling these distant cognitive leaps remain poorly understood. Traditional attention-based interpretability often fails to capture true semantic proximity due to routing artifacts like attention sinks. In this paper, we bypass attention weights to directly analyze the dynamic geometry of the hidden state manifold, p