INDEX

Explanations

point of view

The neuron strongly activates on occurrences of “point of view” (including its preceding modifiers like “from,” “a,” “computational,” “geometric,” etc.).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

</b>

-1.80

-1.66

<i>

-1.38

↵↵

-1.36

 الداخلية

-1.34

 далеко

-1.28

 what

-1.20

𝟹

-1.20

 может

-1.19

 только

-1.16

POSITIVE LOGITS

"}";

1.74

'/',

1.71

 "}";

1.55

 layui

1.52

';

1.48

DataGridView

1.47

 }));

1.42

 "..\..\..\

1.39

}],

1.39

:],

1.39

Activations Density 0.016%