INDEX

Explanations

typing command prompt

The neuron consistently activates on floating‐point numeric tokens (i.e. decimal numbers) in the text.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 hypersur

0.84

 impenetr

0.79

 informativo

0.79

あとは

0.76

 penetrated

0.75

 deterministic

0.74

 drawed

0.74

 extrapolate

0.74

 불구하고

0.74

۰

0.74

POSITIVE LOGITS

and

1.01

in

0.91

im

0.86

ant

0.86

am

0.81

or

0.79

pis

0.79

art

0.77

hans

0.76

our

0.75

Activations Density 0.000%