INDEX

Explanations

intensity and desire

The neuron flags words expressing a speaker’s personal stance or desire (e.g. want, wanted, doubt, envy).

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

ques

-0.87

เยอะ

-0.81

rus

-0.77

 تعدى

-0.76

NoStop

-0.76

 зависимости

-0.73

 Läng

-0.71

ículos

-0.71

ærke

-0.71

Einige

-0.71

POSITIVE LOGITS

 very

3.20

 очень

2.72

 bardzo

2.69

 badly

2.67

 dearly

2.52

 muito

2.36

 дуже

2.33

 rất

2.27

 foarte

2.25

 greatly

2.20

Activations Density 0.058%