INDEX

Explanations

control and responsibility

The neuron detects words and phrases expressing control or the desire for control/power.

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

icals

0.89

 marred

0.88

umuz

0.86

orgetown

0.80

 попа

0.78

kens

0.78

цію

0.77

nels

0.77

oubtedly

0.74

 freshen

0.74

POSITIVE LOGITS

 controlar

1.55

 Kontrolle

1.50

 control

1.48

控制

1.47

 controllo

1.46

 контроль

1.46

 responsibility

1.44

ควบคุม

1.41

 контроля

1.38

 控制

1.38

Activations Density 1.228%