INDEX

Explanations

comparisons and non-english words

This neuron detects the comparative intensifier “more” whenever it appears prominently in text.

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

but

-1.41

But

-1.38

 так

-1.34

and

-1.34

so

-1.31

 will

-1.30

But

-1.22

 Both

-1.21

-1.18

 such

-1.17

POSITIVE LOGITS

或者

1.50

patrón

1.48

 москва

1.38

 zult

1.38

 prezzi

1.38

 jazdy

1.37

обзор

1.36

disfraz

1.35

diseño

1.34

หรือ

1.33

Activations Density 0.021%