categories with years

The neuron strongly activates on entries in the “External links” (and similar reference/link) sections of a page.

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

'',

-1.84

 '+':

-1.63

 ",";

-1.61

^(@

-1.61

chenta

-1.57

carus

-1.55

OLOGIE

-1.54

ᵢ

-1.50

tli

-1.50

[],

-1.49

POSITIVE LOGITS

 will

2.08

1.98

1.85

1.83

1.76

1.75

1.63

1.59

Activations Density 0.152%