INDEX

Explanations

names followed by descriptions or surnames

The neuron is strongly activated by personal proper names (people’s names).

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

например

0.79

 маник

0.77

 váy

0.74

ímetros

0.73

ántica

0.71

 অক্টো

0.71

ímp

0.71

achten

0.70

 сама

0.70

 тема

0.70

POSITIVE LOGITS

 Williams

1.05

 Johnson

0.92

 Sanchez

0.91

 Evans

0.88

 Wilson

0.88

 Cohen

0.87

 Smith

0.86

 Thompson

0.86

 Hernandez

0.86

 Walters

0.85

Activations Density 0.120%