INDEX

Explanations

states or progress

The neuron fires on short declarative phrases that report positive outcomes, progress, or good fortune.

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

-1.58

܆

-1.49

 पढ़

-1.47

 trouva

-1.46

-1.44

 küm

-1.43

chiato

-1.42

EDAD

-1.40

 écou

-1.34

쐈

-1.34

POSITIVE LOGITS

to

2.19

 then

1.95

of

1.85

nipeg

1.52

who

1.52

 потім

1.44

 whereby

1.43

 campeón

1.41

にとっては

1.41

 which

1.37

Activations Density 0.389%