INDEX
Explanations
the term "ic," which appears to denote an attribute or specialized style, potentially in technical or scientific contexts
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
276
+0.12
0.7%
468
+0.12
0.7%
180
+0.11
0.6%
Correlated Neurons
Index
P. Corr.
Cos Sim.
468
+0.12
0.04
395
+0.12
0.04
276
+0.11
0.04
Negative Logits
orderly
-1.59
ľ
-1.42
atement
-1.41
edy
-1.39
Appellants
-1.38
isance
-1.37
arin
-1.35
upon
-1.35
INION
-1.34
©
-1.34
POSITIVE LOGITS
ulated
1.72
works
1.53
lock
1.53
len
1.51
ulation
1.49
ium
1.47
ulate
1.47
lint
1.47
leep
1.47
ethe
1.45
Activations Density 1.031%