INDEX

Explanations

various grammatical particles and prepositions

The neuron fires strongly on Japanese grammatical particles—especially the particle で (and related connective forms like して) that mark adverbial phrases.

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

 that

-2.70

 after

-1.72

if

-1.71

 edisi

-1.70

any

-1.56

 notable

-1.55

 dimmed

-1.52

 just

-1.50

 simple

-1.50

żył

-1.49

POSITIVE LOGITS

this

1.77

 appear

1.60

zeichnete

1.52

 избран

1.51

thanks

1.50

Therefore

1.50

Nevertheless

1.46

 appeler

1.41

 calientes

1.41

 legati

1.40

Activations Density 0.029%