INDEX

Explanations

zero placeholders or default values

The neuron fires strongly on standalone numeric literals (e.g. “0”, “1”, “2”, etc.) 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

涐

-1.34

讥

-1.27

ϰ

-1.26

譏

-1.24

Major

-1.20

之时

-1.20

声を

-1.20

Must

-1.19

Fg

-1.19

-1.14

POSITIVE LOGITS

 since

1.48

 because

1.45

đầm

1.42

alisis

1.38

 perdón

1.32

 this

1.29

 bambú

1.29

ENH

1.27

ը

1.27

if

1.26

Activations Density 0.046%