INDEX

Explanations

need necessity or to perform

The neuron flags words expressing requirement or desire—terms like “нужно,” “necesito,” “quiero,” “прошу,” etc., that signal asking for or needing something.

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

 that

-2.42

 when

-2.27

it

-2.19

 these

-2.19

 this

-2.08

any

-2.06

 there

-1.89

 every

-1.88

 some

-1.85

one

-1.76

POSITIVE LOGITS

to

2.09

1.80

しまって

1.62

ributes

1.62

￡

1.59

to

1.59

<",

1.58

vij

1.56

au

1.55

Activations Density 0.035%