INDEX

Explanations

educational purposes only

The main thing this neuron does is find numerical and quantitative expressions (e.g. multi-digit numbers, years, percentages, and other metrics).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

1.32

 সাধারন

1.25

 loosing

1.23

 পৌছ

1.20

 বেশী

1.19

 তৈরী

1.16

 부터

1.16

 muß

1.13

 skall

1.09

 নির্মান

1.08

POSITIVE LOGITS

ᄒ

1.38

 проє

1.36

ᅲ

1.36

 metaverse

1.34

1.27

🥹

1.27

 Metaverse

1.24

 проєкту

1.23

 macOS

1.22

 onboarding

1.19

Activations Density 0.661%