INDEX

Explanations

positive evaluations following great/good/amazing

The neuron activates on positive evaluative adjectives and other praise words indicating strong positive sentiment.

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

or

-1.59

is

-1.52

to

-1.49

 objedná

-1.43

 não

-1.43

 otorga

-1.43

água

-1.40

 или

-1.38

 أو

-1.34

ही

-1.30

POSITIVE LOGITS

 there

1.70

 people

1.54

 desconocido

1.31

 beautiful

1.29

vânt

1.27

 Certo

1.24

uyen

1.23

舂

1.22

 говорит

1.22

 семья

1.20

Activations Density 0.029%