INDEX

Explanations

As

This neuron strongly activates on the word “as” (i.e. comparative or linking uses of “as”).

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

1.54

м

1.41

truck

1.24

데요

1.23

tion

1.22

데

1.22

ERE

1.21

’

1.21

the

1.18

т

1.17

POSITIVE LOGITS

ymmet

1.85

sembles

1.81

ynchron

1.63

inine

1.63

ignment

1.34

cribable

1.27

cribing

1.26

ynchronously

1.20

 минимум

1.18

itent

1.18

Activations Density 0.360%