INDEX

Explanations

continue reading

The neuron consistently fires on blog‐style “Continue reading”/“Reading” navigation prompts and on numeric pagination or citation markers (e.g. page numbers and float values).

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

‘

-1.58

 розо

-1.54

вето

-1.45

趼

-1.43

тель

-1.43

)}}

-1.42

 latter

-1.41

 видимо

-1.40

’

-1.40

 What

-1.40

POSITIVE LOGITS

捱

1.75

 theses

1.63

 capita

1.55

,,,

1.48

 quando

1.46

 then

1.45

¹,

1.45

킁

1.43

 ezek

1.42

れば

1.41

Activations Density 0.003%