INDEX

Explanations

data assignments and definitions

The neuron strongly activates on occurrences of the identifier “data” (e.g. variables or array keys named data) in code.

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

or

-2.20

in

-2.00

 with

-1.96

as

-1.91

 after

-1.76

for

-1.76

ueroa

-1.75

 since

-1.59

on

-1.58

 from

-1.55

POSITIVE LOGITS

 frança

1.61

 semblent

1.54

 bornes

1.53

 restant

1.51

1.50

 bricola

1.48

雠

1.48

 staffel

1.48

frist

1.48

都沒

1.48

Activations Density 0.002%