INDEX

Explanations

things following 'super' or 'sub'

The neuron detects the first part of compound noun phrases (e.g. “snow” in snowshoes, “sky” in skydivers, “camper” in camper vans, “mother” in motherboards, 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

if

-1.83

 Even

-1.55

as

-1.54

 numerous

-1.42

one

-1.42

 even

-1.41

 ingenious

-1.40

 using

-1.39

POSITIVE LOGITS

 новые

1.55

ਗ਼

1.50

 diſt

1.41

罍

1.41

jelaskan

1.41

好久

1.40

membre

1.38

 réellement

1.36

ны

1.36

閖

1.34

Activations Density 0.292%