INDEX

Explanations

intensifier modifying descriptor

This neuron fires on degree-and-intensity words—adverbs and adjectives that express how strong or extreme something is (e.g. “very,” “pretty,” “bleak,” “complicated”).

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

睽

-1.55

是一种

-1.53

 Weltkrieg

-1.44

those

-1.39

these

-1.34

kah

-1.33

would

-1.32

many

-1.32

駅の

-1.32

 sorun

-1.30

POSITIVE LOGITS

 into

1.35

 speziellen

1.27

in

1.25

ícios

1.23

 during

1.21

 розовый

1.18

*}[

1.17

 osób

1.16

指定

1.16

 speziell

1.15

Activations Density 0.051%