INDEX

Explanations

morning today book discovered far inside around paper

The neuron lights up on semantically “contentful” tokens—especially proper names, headline nouns, prices, dates/times, and other key informational 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

by

-2.30

or

-1.98

 like

-1.95

for

-1.76

 there

-1.57

-1.48

 some

-1.48

 until

-1.47

 zweck

-1.45

 with

-1.39

POSITIVE LOGITS

也很

1.70

ियाँ

1.42

ぱ

1.34

ぽ

1.30

 także

1.26

ément

1.23

 substan

1.23

wnież

1.22

}\

1.21

だけでなく

1.20

Activations Density 0.121%