INDEX

Explanations

lists and punctuation

The neuron is a high‐level content detector that lights up on open‐class words—especially nouns and verbs that carry concrete descriptive or action meaning—while ignoring most function words.

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

to

-2.02

 beiden

-1.70

 COMPLE

-1.53

 hvil

-1.50

 seksi

-1.48

 osal

-1.45

 sekä

-1.45

‪.‬‬

-1.45

沅

-1.44

“…

-1.42

POSITIVE LOGITS



1.64

dáms

1.63

zdroj

1.60

for

1.46

 like

1.44

encana

1.39

 jaula

1.38

 zeigte

1.38

 рассказывает

1.38

 most

1.36

Activations Density 0.097%