INDEX

Explanations

number 2 followed by period

The neuron fires on sentence-initial discourse cues—words like “So,” “Now,” “Since,” “With,” “Turns (out),” etc., that mark transitions or topic shifts.

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

 explanations

0.76

 explanation

0.64

 waters

0.55

 deliberations

0.54

 motives

0.54

 significant

0.53

 justifications

0.52

 discussions

0.52

erdere

0.52

 писал

0.51

POSITIVE LOGITS

Enter

1.80

 Enter

1.74

 enter

1.53

Introducing

1.53

 Introducing

1.44

enter

1.38

ENTER

1.38

 ENTER

1.38

 enters

1.30

introducing

1.30

Activations Density 0.328%