INDEX

Explanations

conditional statements and code definitions

This neuron fires on the initial, capitalized directive words (e.g. Get, Check, If, This) that begin code comment lines.

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



-1.48

CESO

-1.38

 aéreas

-1.32

 griechischen

-1.32

žno

-1.31

melidir

-1.28

 prohibido

-1.26

 conocidas

-1.23

 poderoso

-1.22

 inalámbrica

-1.22

POSITIVE LOGITS

you

1.41

to

1.27

of

1.27

岔

1.23

1.22

 which

1.21

times

1.20

by

1.20

idase

1.15

Cited

1.09

Activations Density 0.071%