INDEX

Explanations

set to use

The neuron activates on instances of the verb “set” (often in “set … to …”) or closely related imperative configuration/assignment phrases.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

而是

-1.48

$<

-1.41

MERS

-1.37

-1.27

שמי

-1.27

 ألمانيا

-1.25

wou

-1.23

itemize

-1.23

seamnă

-1.23

TERS

-1.23

POSITIVE LOGITS

 about

1.94

 says

1.38

 regarding

1.37

has

1.33

してた

1.27

as

1.27

 også

1.24

最悪

1.23

 concerning

1.23

Activations Density 0.066%