INDEX

Explanations

able to or incapable of

The neuron fires strongly on positive, descriptive adjectives (especially evaluative or promotional qualifiers).

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

 andra

0.79

 అనంతరం

0.79

 discusión

0.78

 другие

0.74

 diğer

0.72

 wilayah

0.71

 conclusiones

0.71

 actuales

0.70

 heridas

0.70

 antiguas

0.69

POSITIVE LOGITS

 adaptable

0.89

 contains

0.82

 personable

0.82

 able

0.77

contains

0.77

 suited

0.76

Contains

0.76

 multifunctional

0.75

 accessible

0.74

readable

0.73

Activations Density 0.107%