INDEX

Explanations

recommendations and broad audiences

The neuron detects endorsement statements—phrases where the speaker “recommends” something to a person or group.

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

鐧

-1.27

 recognisable

-1.26

Otros

-1.26

ICIAS

-1.23

埵

-1.23

liziert

-1.23

鈊

-1.22

Hermoso

-1.20

惭

-1.20

 Bereichen

-1.19

POSITIVE LOGITS

 because

1.61

^{-

1.50

1.48

."

1.46

".

1.43

1.41

 family

1.40

();

1.38

.”

1.35

:\/\/

1.35

Activations Density 0.002%