INDEX

Explanations

notable figures and entities

The neuron fires on evaluative adjectives ending in “-ous,” notably words like “prestigious” or “infamous.”

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

 interessante

0.73

 interesting

0.73

 চমৎকার

0.64

꽤

0.63

很好的

0.62

 miserable

0.59

奇怪

0.59

 intéressante

0.59

 interesante

0.57

interesting

0.57

POSITIVE LOGITS

्स

0.83

 زمانہ

0.67

 Denison

0.59

DOUBLE

0.59

 Double

0.57

 Jameson

0.57

 Defts

0.57

 erster

0.56

 Bourd

0.56

 نیز

0.55

Activations Density 0.007%