INDEX

Explanations

verbs followed by specific modifiers

This neuron fires on the first substantial (content) word after a sentence or clause boundary, marking sentence-initial nouns or verbs.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 itself

1.59

 βρίσκεται

1.50

 ತನ್ನ

1.49

was

1.45

 ఉంది

1.38

它的

1.38

its

1.37

 ہے

1.30

is

1.27

 wasn

1.27

POSITIVE LOGITS

 đều

2.67

都很

2.59

 themselves

2.56

 eivät

2.39

są

2.33

 هستند

2.28

都是

2.25

 were

2.23

were

2.23

 vannak

2.23

Activations Density 0.562%