INDEX

Explanations

The neuron activates on numeric literals (integer sequences) in the text.

keywords and code directives

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

 goûts

-1.20

 défis

-1.16

-1.12

NEN

-1.12

 probablement

-1.12

 suppos

-1.11

 diriger

-1.10

を見ながら

-1.09

 fekete

-1.08

 Ostro

-1.08

POSITIVE LOGITS

and

1.88

all

1.77

if

1.60

 only

1.46

so

1.41

 before

1.32

at

1.28

ALL

1.23

 when

1.23

any

1.20

Activations Density 0.001%