INDEX

Explanations

dimensions, counts, and programming

The neuron responds to occurrences of length/size queries or properties (e.g. calls like strlen(s) or use of .length/.size).

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

 either

-1.13

 only

-1.11

or

-1.05

 later

-1.02

૭

-1.00

 week

-0.95

 visit

-0.90

 please

-0.89

 like

-0.87

 supports

-0.87

POSITIVE LOGITS

 fotball

1.07

១

1.03

<0xA3>

1.02

"";

0.98

 enthousi

0.98

 adverti

0.95

();

0.94

 monstru

0.93

 manifeste

0.93

 tasar

0.91

Activations Density 0.002%