Skip to content

Confirm magic_proxy_phase_R regression (0.864→0.778) in the alpha=0.12 stack isn't trading away dream-magic #332

Description

@NickFlach

Source: note from closed PR #281; relates to the "magic gives it gravity" research thread.

In the alpha=0.12 engine_a stack that improved transfer/fitness, magic_proxy_phase_R regressed 0.8643 → 0.7785 (−0.086). magic_proxy_phase_R is the quantum-info "magic"/non-Clifford content proxy associated with dreaming (vs recall).

Action: confirm the transfer/fitness win isn't silently trading away the dream "magic" signal. If the regression is real and load-bearing, it should be a tracked cost in the fitness function rather than silent drift.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions