INDEX

Explanations

which item or kind

The neuron consistently fires on occurrences of the interrogative word “which.”

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

﹍﹍

-1.40

怎么回事

-1.30

EDAD

-1.30

 aguj

-1.29

理由は

-1.27

oused

-1.23

hamshire

-1.22

matical

-1.22

,‎

-1.21

antry

-1.21

POSITIVE LOGITS

of

4.22

one

2.48

is

2.03

 ones

1.88

are

1.72

 kind

1.70

 type

1.67

 part

1.66

of

1.58

 version

1.54

Activations Density 0.021%