INDEX

Explanations

contractions and possessives

The neuron fires on single-quote/apostrophe characters (e.g. in contractions or quoted speech).

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

Dzień

-1.82

 escritório

-1.61

Čes

-1.59

Kiedy

-1.59

Wię

-1.57

 délicieux

-1.55

Czym

-1.55

Estás

-1.55

 ترجمه

-1.54

〣

-1.52

POSITIVE LOGITS

and

2.06

1.75

 they

1.59

1.51

 Kalt

1.50

='

1.49

 when

1.47

},

1.46

as

1.45

 both

1.45

Activations Density 0.023%