INDEX

Explanations

announcements and releases

This neuron responds to auxiliary‐verb sequences expressing modality or future action (e.g. “seems to,” “will be releasing,” “start to”).

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

 plakat

-1.01

 bluz

-0.99

 kolon

-0.98

Klik

-0.96

 demandé

-0.96

RestTemplate

-0.93

 karton

-0.93

 использованием

-0.93

 tambahan

-0.92

 tunik

-0.92

POSITIVE LOGITS

 releasing

1.95

 release

1.86

 released

1.70

 releases

1.64

 introduce

1.48

 introducing

1.44

 introduces

1.38

 wprowad

1.34

 introduced

1.34

 update

1.33

Activations Density 0.083%