mello-scale-k-values-cf-t
IN premise — summaries/2026/08/24/zhong-2023-mquake-s5-mello-a-proposal-for-editing-large.md
Created 2026-08-25T02:59:06+00:00
MeLLo scale evaluation uses k ∈ {1, 100, 1000, 3000} simultaneous edits on MQuAKE-CF and k ∈ {1, 100, 500, 1868} on MQuAKE-T.
Summary
This documents the specific batch sizes used when testing the MeLLo system under load, measured by how many edits are applied at once across two benchmark scenarios. It matters because the maximum batch size differs between the two scenarios (3000 versus 1868), meaning the evaluation is scaled to each dataset's actual capacity rather than using a one-size-fits-all number, and any comparison of results must account for those different operating points.