spb/wpai0001_uqo

1% source: experiments/summaries/tier1_taskvector_e105e3d91386/results.json2\begin{tabular}{llcccccc}3\toprule4Model & task & zero-shot & few-shot & best layer & vector added (zero-shot) & random vector & few-shot minus direction \\5\midrule6llama3b (28 layers, $d=3072$) & antonym & 0.16 & 0.88 & L14 & 0.84 & 0.06 & 0.91 \\7llama3b (28 layers, $d=3072$) & capital & 0.00 & 1.00 & L14 & 1.00 & 0.00 & 1.00 \\8llama3b (28 layers, $d=3072$) & morph & 0.00 & 0.35 & L2 & 0.00 & 0.00 & 0.35 \\9llama3b (28 layers, $d=3072$) & addthree & 0.00 & 1.00 & L14 & 0.80 & 0.00 & 1.00 \\10qwen1.5b (28 layers, $d=1536$) & antonym & 0.22 & 1.00 & L16 & 0.88 & 0.47 & 1.00 \\11qwen1.5b (28 layers, $d=1536$) & capital & 0.00 & 1.00 & L22 & 0.86 & 0.00 & 1.00 \\12qwen1.5b (28 layers, $d=1536$) & morph & 0.00 & 0.33 & L2 & 0.00 & 0.00 & 0.33 \\13qwen1.5b (28 layers, $d=1536$) & addthree & 0.00 & 0.86 & L2 & 0.00 & 0.00 & 0.86 \\14qwen3b (36 layers, $d=2048$) & antonym & 0.25 & 0.94 & L25 & 0.88 & 0.06 & 0.97 \\15qwen3b (36 layers, $d=2048$) & capital & 0.00 & 1.00 & L25 & 0.82 & 0.00 & 1.00 \\16qwen3b (36 layers, $d=2048$) & morph & 0.00 & 0.40 & L2 & 0.00 & 0.00 & 0.40 \\17qwen3b (36 layers, $d=2048$) & addthree & 0.00 & 1.00 & L2 & 0.00 & 0.00 & 1.00 \\18\bottomrule19\end{tabular}20