A/B verdict tool for recorded optimization-run ledgers. Use when the user wants two (or more) runs of the GPU-optimization loop compared and judged: knowledge-base ON vs OFF ablations, "which config/side won?", "did KB actually help — real effect or noise?", round-by-round or convergence-speed analy
A/B verdict tool for recorded optimization-run ledgers. Use when the user wants two (or more) runs of the GPU-optimization loop compared and judged: knowledge-base ON vs OFF ablations, "which config/side won?", "did KB actually help — real effect or noise?", round-by-round or convergence-speed analysis, or a sweep summary across several run pairs. Trigger equally on Korean phrasing (온오프 비교, 어느 쪽이
This entry publishes no npm package, so Forge has no dependency tree for it. That is a gap in coverage — not a statement that it has no dependencies.