INDEX

Explanations

quoted concepts or labels

This neuron activates on words that immediately follow a quoted term or phrase.

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

骉

-1.21

 Promotes

-1.12

 disagrees

-1.12

 reaff

-1.10

 proclaims

-1.05

 peculiarly

-1.04

लिखित

-1.04

Ꭾ

-1.03

 trône

-1.03

问道

-1.03

POSITIVE LOGITS

 that

1.64

for

1.62

 when

1.56

 which

1.52

or

1.30

any

1.29

 their

1.25

as

1.23

 before

1.23

 this

1.21

Activations Density 0.164%