preprint
Open access
Microscopic Pe: Measuring Péclet Flow on Circuit Attribution Graphs
Research footprint
At a glance
- Citations
- 0
- References
- 0
- Comments
- 0
Paper overview
Abstract
Anthropic's circuit tracing (2025) — attribution graphs revealing computational paths through language models — provides the microscope for measuring Péclet number at the feature level. We define Pe on attribution graph edges as the ratio of directed information flow (drift) to undirected spreading (diffusion) through cross-layer transcoder features. Jailbreak circuits should exhibit measurable Pe gradients: Pe increases along the computational path. The 12 jailbreak patterns detected by the Void Framework's Twilight monitoring should correspond to specific high-Pe subgraphs. This connects macroscopic Pe scoring (N=1,344 platforms) to microscopic mechanism, bridging mechanistic interpretability and thermodynamic field theory.
Record transparency
Publication details
- DOI
- 10.5281/zenodo.19261142
- OpenAlex
- W7141325159
- Document type
- preprint
- Language
- EN
- Source
- Zenodo (CERN European Organization for Nuclear Research)
- Last metadata update
Comments
Log in to join the discussion.