INDEX

Explanations

The neuron flags expressions of desire or intention—verbs like want, wants, wish, and looking to (e.g. “wants the feel,” “want a device,” “wish to have,” “looking to get”).

wanting to gain or achieve

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

if

-1.37

 Dacă

-1.30

 your

-1.20

you

-1.20

 даже

-1.16

 حتی

-1.16

Și

-1.16

 stwier

-1.14

 статье

-1.12

it

-1.10

POSITIVE LOGITS

 their

2.20

他们

1.47

who

1.45

themselves

1.39

 ihrem

1.34

or

1.31

他們

1.28

 ihrer

1.23

 themselves

1.22

 mereka

1.19

Activations Density 0.037%