wassname 34a2eec704 viz: floor->ceiling as two normalized panels (best vs control vs reference) ...

Rework per feedback: hack and solve are not opposites, so they get separate
floor->ceiling axes (each 0=floor..1=ceiling) rather than sharing a zero -- this
also stops solve (range ~0.13-0.22) being squished next to hack (0-0.61).
Minimal: routeV per-token (best) vs random-V (direction control) vs the SGTM
gradient-routing paper placed on the same floor->ceiling % axis (approx, LM task).

Reads: hack suppression 93% best / 84% control / ~98% reference (9pp = direction
signal); solve gained +17% / -17% / ~95% (far from ceiling -- model barely learns
to solve in 60 steps). Moved out/plots -> out/figs.

Co-Authored-By: Claudypoo <288921227+claudypoo@users.noreply.github.com>

2026-06-09 09:55:03 +00:00

84 KiB

1641x475px

Raw History