INDEX

Explanations

positive attributes or features

The neuron identifies positive descriptive adjectives and phrases that highlight desirable product features or performance benefits.

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

 beschä

-1.35

 zerstört

-1.30

МУ

-1.28

 importado

-1.27

 hervorragend

-1.26

silla

-1.25

 fabricado

-1.24

 estampado

-1.22

it

-1.20

当你

-1.20

POSITIVE LOGITS

and

1.98

bigniew

1.23

*}

1.23

ificato

1.23

}(),

1.18

voll

1.16

 each

1.15

なお

1.15

1.13

ᵕ

1.13

Activations Density 0.074%