INDEX

Explanations

@ followed by name or command

The neuron is picking out the “@” symbol that marks social‐media handles or user mentions.

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

<bos>

-1.95

-1.81

-1.67

IANGLES

-1.67

 acompañada

-1.55

了一眼

-1.53

``

-1.52

着一

-1.51

ichteten

-1.51

POSITIVE LOGITS

all

1.77

the

1.72

にします

1.63

Beskrivning

1.56

 each

1.54

 have

1.50

him

1.45

あらゆる

1.43

 their

1.41

1.37

Activations Density 0.004%