INDEX

Explanations

x followed by hex characters

The neuron specializes in spotting numeric literal tokens (including decimal and hexadecimal 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

 that

-1.66

any

-1.58

 what

-1.53

Having

-1.46

Even

-1.46

all

-1.45

Have

-1.42

 with

-1.40

 have

-1.38

With

-1.38

POSITIVE LOGITS

 deliciously

1.34

 horrid

1.31

™

1.30

 delightfully

1.23

 ridiculously

1.18

bh

1.18

露出了

1.13

︲

1.13

Uf

1.13

 💕

1.13

Activations Density 0.006%