INDEX

Explanations

By followed by pronouns

The neuron detects the key topical word that kicks off a new clause or section—typically the main noun, verb, or date in prepositional/time phrases (e.g. “browsing,” “email,” “college,” “free[time],” “tutorial,” or a year).

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

 there

-2.16

his

-1.89

 that

-1.74

 their

-1.61

There

-1.50

it

-1.47

-1.37

),

-1.34

 this

-1.30

 such

-1.30

POSITIVE LOGITS

1.98

 gouttes

1.23

 réve

1.20

 imprimer

1.12

 उन्होंने

1.11

 unaweza

1.09

 nimmer

1.09

you

1.09

she

1.05

țin

1.05

Activations Density 0.158%