INDEX

Explanations

in addition to

The neuron primarily detects numeric tokens (e.g. standalone numbers, years, counts, or measurements).

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

 both

-1.48

 zarówno

-1.40

one

-1.35

 навіть

-1.23

 after

-1.23

 estekak

-1.18

 paesaggio

-1.13

 ऐसा

-1.11

 only

-1.10

 first

-1.09

POSITIVE LOGITS

 being

2.53

the

1.69

being

1.55

通常の

1.43

 oczy

1.39

 étant

1.38

 obvious

1.36

 일반

1.32

its

1.30

 להיות

1.30

Activations Density 0.034%