INDEX

Explanations

execution and killing

The neuron fires on any form of “execute” (and its derivatives like execution, executed, executioner, etc.), i.e. words referring to carrying out or undergoing capital punishment.

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

().

-1.90

 mischie

-1.83

がない

-1.66

 партнер

-1.62

或者是

-1.62

不是很

-1.58

whom

-1.58

裡的

-1.55

(),

-1.55

 anderen

-1.54

POSITIVE LOGITS

try

1.74

 what

1.73

欵

1.64

 make

1.63

by

1.50

灕

1.50

 umfang

1.48

at

1.46

an

1.44

“

1.42

Activations Density 0.011%