INDEX

Explanations

The neuron activates on words that express enabling or facilitating actions (e.g., “allow,” “help,” “aid,” “lead”).

enabling or causing outcomes

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

 after

-1.71

if

-1.66

 despite

-1.37

 while

-1.33

 from

-1.27

 There

-1.27

 during

-1.26

on

-1.24

or

-1.23

 Despite

-1.22

POSITIVE LOGITS

 tejto

1.46

 caminhada

1.45

 této

1.36

 boneca

1.32

 dáms

1.28

 både

1.28

 estadia

1.26

 olacaktır

1.23

 nuvem

1.23

 identidade

1.23

Activations Density 0.091%