INDEX

Explanations

boolean values

The neuron activates on numeric constants (integer or floating‐point literals) in code.

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

of

-2.42

to

-1.42

 from

-1.34

 that

-1.32

 gestern

-1.27

ar

-1.23

Aho

-1.21

となり

-1.20

 özellikle

-1.16

 muhte

-1.16

POSITIVE LOGITS

………

1.52

This

1.49

<0xE6>

1.42

Ｒ

1.41

cuerdo

1.40

…….

1.39

othesis

1.39

 deudas

1.34

But

1.30

but

1.27

Activations Density 0.008%