INDEX

Explanations

people feel

The neuron fires on numeric tokens (numbers, measurements, and numeric codes).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

Fic

0.63

 mostly

0.54

 Latest

0.53

 Mostly

0.52

 ближе

0.52

 Sharp

0.51

更多

0.51

 दिनेश

0.50

ODM

0.50

仍在

0.50

POSITIVE LOGITS

omer

0.58

pir

0.57

 sign

0.56

obatan

0.56

ondere

0.55

ldecode

0.55

Sign

0.55

二维

0.55

sign

0.55

aint

0.54

Activations Density 0.000%