INDEX

Explanations

operated by

The neuron detects statements of institutional operation—especially the phrase “operated by” and its surrounding words.

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

 they

-2.22

-1.92

 that

-1.73

 problemer

-1.54

ágenes

-1.54

 hän

-1.54

 påvir

-1.52

全体の

-1.51

riezmann

-1.46

葑

-1.46

POSITIVE LOGITS

 This

1.98

1.86

–

1.86

 questo

1.72

 notwendig

1.63

1.62

 один

1.59

 Many

1.59

 dette

1.58

 werden

1.57

Activations Density 0.039%