INDEX
Explanations
specific numerical values and their associated contexts, including citations and mathematical operations
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
342
+0.17
1.0%
153
+0.13
0.8%
276
+0.12
0.7%
Correlated Neurons
Index
P. Corr.
Cos Sim.
111
+0.17
0.21
71
+0.13
0.22
156
+0.12
0.22
Negative Logits
ör
-1.62
Editor
-1.57
better
-1.47
uities
-1.39
icz
-1.37
Include
-1.35
umab
-1.35
umi
-1.34
anium
-1.34
wald
-1.34
POSITIVE LOGITS
»¿
3.36
º
3.32
¾
3.23
¶
3.21
Ĥ¬
3.16
ģ
3.13
»
3.09
¿
3.08
ij
2.99
¬
2.99
Activations Density 3.775%