INDEX

Explanations

numbers and punctuation

The neuron fires on numeric tokens and list‐item markers (e.g. step numbers, section numbers, dates, and other numerals).

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

＊＊＊＊

-1.55

 стоковые

-1.53

﹕

-1.48

 МЕ

-1.47

bestimm

-1.47

鬭

-1.45

︰

-1.43

 forno

-1.41

%");

-1.40

 meni

-1.38

POSITIVE LOGITS

is

1.80

 then

1.61

がある

1.56

があり

1.54

1.45

 about

1.43

you

1.41

 come

1.40

↵↵

1.38

 others

1.34

Activations Density 0.008%