Determine whether the coordinate intervention reveals a general cost of coordinate fields

Determine whether removing structured recipe-level coordinate fields from agentic text memory produces a general performance cost, rather than only the unresolved interaction observed in the coordinate de-prescription screen.

Background

The coordinate study tested only a limited two-cell directional screen and did not establish the predicted interaction. Because the intervention removed structured coordinate fields while leaving narrative summaries intact, the experiment does not resolve whether coordinate information is generally beneficial or harmful in the live agentic harness.

References

The intervention removed structured recipe-level coordinate fields only and left narrative summaries unchanged, so this study does not close the coordinate-specific mechanism or establish a general cost of removing coordinate fields.

— Beyond Success Rate: Interventional Evaluation of Memory in Vision-Language-Action Systems  (2610.23372 - Luo et al., 16 Oct 2026) in Appendix B, Section “Coordinate De-Prescription”