INDEX

Explanations

all

The neuron is looking for the word “all” (especially when it appears as the quantifier “all …”).

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

 miei

-1.51

鹛

-1.37

äischen

-1.37

 mismas

-1.36

ilmente

-1.35

还会

-1.31

CARACTER

-1.30

</i>

-1.30

zelf

-1.30

ackerel

-1.29

POSITIVE LOGITS

 from

1.91

 betroffen

1.29

 previous

1.27

šení

1.23

but

1.23

it

1.23

1.16

 previously

1.16

the

1.15

which

1.15

Activations Density 0.052%