INDEX

Explanations

script, functional, array, User, yes, drive

The neuron strongly activates on programming‐related identifiers and technical terms (e.g. data‐structure names like “array,” function names like “length” or “script,” and similar code tokens).

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

茈

-1.59

 Etats

-1.38

 appré

-1.30

This

-1.27

 venir

-1.26

 obé

-1.25

 "]";

-1.25

 Batalla

-1.24

 donne

-1.23

 réessayer

-1.23

POSITIVE LOGITS

 because

1.98

for

1.76

 when

1.73

by

1.45

and

1.43

 extinción

1.43

 first

1.41

 greatest

1.39

 normally

1.38

 according

1.34

Activations Density 0.048%