INDEX

Explanations

The neuron detects standalone numeric constants in code (e.g. array sizes or literal numbers).

SQL syntax

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

んですよね

-1.41

 Therefore

-1.39

こともあり

-1.36

為に

-1.29

our

-1.27

 Pří

-1.26

 Drž

-1.24

づけ

-1.23

 such

-1.23

kenal

-1.21

POSITIVE LOGITS

を

1.41

は

1.34

薔薇

1.29

ですが

1.27

 naves

1.26

 equivalente

1.23

では

1.23

 anormal

1.22

などの

1.20

が

1.19

Activations Density 0.004%