INDEX

Explanations

each or different items

The neuron fires on quantitative and enumeration language—tokens that introduce or refer to numbers, measurements, or counts (e.g. “each,” “different,” “selected,” “large,” “single,” etc.).

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

in

-1.95

not

-1.66

 with

-1.62

 first

-1.53

and

-1.53

two

-1.44

-1.41

 descrit

-1.41

 probably

-1.38

for

-1.37

POSITIVE LOGITS

those

1.53

its

1.52

 determinados

1.42

 variados

1.38



1.35

部分が

1.32

 choisi

1.31

diensten

1.28

 hordes

1.28

 choisir

1.28

Activations Density 0.188%