INDEX

Explanations

end of code blocks

The neuron detects numeric literals (integers and floating‐point numbers) 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

厹

-1.34

伅

-1.34

 drôle

-1.33

 essayé

-1.30

ichtete

-1.28

 itinéraire

-1.27

⼒

-1.26

previously

-1.26

 reklama

-1.25

yym

-1.24

POSITIVE LOGITS

1.54

1.48

1.45

1.41

any

1.41

1.32

}{

1.27

ńskim

1.23

 when

1.23

ers

1.22

Activations Density 0.005%