INDEX

Explanations

like, may, device commands

This neuron detects direct second‐person address, i.e. occurrences of “you.”

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

<0x98>

0.79

יל

0.74

ਤੇ

0.71

ברי

0.70

ت

0.68

𝙈

0.68

elry

0.67

<0x93>

0.65

𝙼

0.65

ک

0.64

POSITIVE LOGITS

 chamar

0.86

 cultivos

0.84

 Beit

0.82

 mecanismos

0.80

 tokamaks

0.80

 thuring

0.80

 grabs

0.79

 barometer

0.79

 Maduro

0.78

 Sinal

0.78

Activations Density 0.002%