INDEX

Explanations

isolate and manipulate

This neuron detects numeric values (especially numbers, often with decimals) in the text.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 هستند

1.05

 hauptsächlich

1.05

 nejsou

1.00

 estavam

0.98

 molte

0.96

 бывают

0.96

 głównie

0.93

 esistono

0.90

 gehad

0.89

 estaban

0.88

POSITIVE LOGITS

 then

1.34

将其

1.30

把它

1.30

then

1.24

Then

1.09

delete

1.04

Kemudian

1.03

 remove

1.02

 delete

1.02

然后

1.02

Activations Density 0.801%