INDEX

Explanations

converting to new states

The neuron lights up on salient content words—especially thematic nouns and key verbs that carry the main topic of a sentence.

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

of

-1.46

is

-1.39

but

-1.37

<u>

-1.23

 similarities

-1.20

 tetapi

-1.20

 such

-1.19

<eos>

-1.18

but

-1.18

 recent

-1.14

POSITIVE LOGITS

 when

1.84

みなさま

1.39

蒾

1.38

剮

1.36

entière

1.35

Ᏹ

1.34

皆さま

1.31

︓

1.30

toolStripButton

1.30

 chyb

1.30

Activations Density 0.069%