mirror of
https://github.com/hanzoai/papers.git
synced 2026-10-29 07:34:31 +00:00
zooqueen
67816f32ac
The engine's ablation on all ten conversations, the test split run once after the dev freeze, the same rows over all-MiniLM-L6-v2 at k=10, and the frozen configuration by k are generated by hanzoai/cloud bench/brain/tables.mjs. LoCoMo-Conv is unreleased, so that section now says so and reports the protocol check on the original questions instead of a number nobody can reproduce.
14 lines
529 B
TeX
14 lines
529 B
TeX
% generated by hanzoai/cloud bench/brain/tables.mjs — the frozen engine by k; a number quoted from here carries its k
|
|
\begin{tabular}{lrrr}
|
|
\toprule
|
|
frozen configuration & $k=5$ & $k=10$ & $k=20$ \\
|
|
\midrule
|
|
test split, multi-hop ALL & 15.9 & 26.4 & 37.5 \\
|
|
test split, multi-hop ANY & 69.2 & 78.8 & 88.0 \\
|
|
test split, single-hop ALL & 74.1 & 81.7 & 87.7 \\
|
|
all ten, multi-hop ALL & 17.4 & 27.0 & 39.0 \\
|
|
all ten, multi-hop ANY & 68.8 & 79.8 & 87.9 \\
|
|
all ten, single-hop ALL & 72.7 & 80.4 & 86.8 \\
|
|
\bottomrule
|
|
\end{tabular}
|