INDEX

Explanations

calculated or modified data

The neuron is triggered by words that start with a capital letter (i.e. capitalized tokens).

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

 strides

1.20

ногда

1.19

 claimants

1.18

 recollect

1.17

倆

1.17

SHO

1.16

 alignItems

1.15

くちゃ

1.15

вому

1.14

 addicts

1.14

POSITIVE LOGITS

ependent

1.73

ifferent

1.70

etection

1.66

ierten

1.53

્સ

1.48

by

1.39

omin

1.30

تي

1.30

 volna

1.27

ال

1.26

Activations Density 10.204%