Files
papers/tables/locomo-minilm.tex
zooqueen 67816f32ac Paper 1 reads its retrieval tables from the runs
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.
2026-09-09 03:07:37 -07:00

17 lines
952 B
TeX

% generated by hanzoai/cloud bench/brain/tables.mjs — all-MiniLM-L6-v2 vectors, k=10, LoCoMo-Conv protocol embedder
\begin{tabular}{lrrrrrrrrrr}
\toprule
& \multicolumn{2}{c}{multi-hop} & single & temporal & open & all & & & pool & p50 \\
\cmidrule(lr){2-3}
Configuration & ALL@10 & ANY@10 & ALL@10 & ALL@10 & ALL@10 & R@10 & nDCG & supported & cands & ms \\
\midrule
cosine over turns & 9.6 & 49.6 & 51.4 & 46.1 & 16.3 & 45.3 & 31.1 & 53.1 & 40 & 0.25 \\
+ BM25 & 13.8 & 58.9 & 67.2 & 62.0 & 23.9 & 59.2 & 43.7 & 67.6 & 54 & 0.31 \\
+ atomic fact index & 20.9 & 73.4 & 73.1 & 72.0 & 25.0 & 66.7 & 52.0 & 75.6 & 73 & 0.42 \\
+ adjacency & 21.6 & 74.5 & 74.6 & 72.6 & 27.2 & 67.7 & 53.3 & 76.3 & 77 & 0.42 \\
+ second hop from resolved state & 24.1 & 78.0 & 77.5 & 73.5 & 27.2 & 70.1 & 55.8 & 78.5 & 78 & 0.76 \\
\midrule
global reciprocal-rank fusion & 20.2 & 74.8 & 73.4 & 74.1 & 28.3 & 67.2 & 48.6 & 76.1 & 143 & 0.43 \\
\bottomrule
\end{tabular}