INDEX

Explanations

command flags and system calls

The neuron is chiefly triggered by command‐line option tokens (i.e. dash-prefixed compiler or tool flags).

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

by

-2.16

as

-1.93

and

-1.74

be

-1.73

出了一

-1.72

in

-1.70

来越

-1.56

’

-1.55

“

-1.55

–

-1.46

POSITIVE LOGITS

 таких

1.84

 elementów

1.77

 niñas

1.72

 reason

1.72

↵↵

1.71

 tecido

1.68

這種

1.68



1.66

Selama

1.65

Semua

1.64

Activations Density 0.032%