INDEX

Explanations

error and string literals

The neuron activates on numeric literals (integer or floating‐point numbers) in source 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

-1.37

-1.34

蓥

-1.33

 функциони

-1.31

first

-1.30

Ö

-1.28

 véhic

-1.27

li

-1.27

紧接着

-1.27

POSITIVE LOGITS

 such

1.85

 this

1.83

เพราะ

1.45

 着用

1.37

 after

1.33

前は

1.32

 Towel

1.30

aları

1.28

گل

1.28

ümüş

1.27

Activations Density 0.002%