INDEX

Explanations

turning up or down

The neuron activates on words and phrases that denote adjusting or changing a level or intensity—e.g. “turn up,” “crank,” “ramp up,” “increase,” “decrease,” “ratchet up,” “turn down.”

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

 Thus

-1.85

 дов

-1.79

 смог

-1.70

 Notably

-1.52

 født

-1.48

 Nowadays

-1.48

 Roughly

-1.47

 знал

-1.46

 Lastly

-1.45

POSITIVE LOGITS

the

2.22

is

1.54

1.52

 then

1.51

 rozwój

1.43

 laquelle

1.40

 pomysł

1.38

↵

1.38

<i>

1.36

 sprzedaż

1.35

Activations Density 0.019%