INDEX

Explanations

manner, degree, frequency

The neuron specializes in spotting degree- or frequency-modifying adverbs and qualifiers (e.g. typically, always, partially, precisely).

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

künfte

-1.38

здрав

-1.33

läufer

-1.29

cartes

-1.26

 últimas

-1.26

verschluss

-1.25

"=>$

-1.23

Toledo

-1.23

"},

-1.23

'},

-1.21

POSITIVE LOGITS

and

1.78

When

1.45

 which

1.38

Following

1.31

 within

1.29

 only

1.23

which

1.23

The

1.22

 connected

1.22

 will

1.20

Activations Density 0.253%