INDEX

Explanations

waking and alertness

The main thing this neuron does is detect mentions of “wake up” (including forms like wake, woke, or woken followed by up).

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

↵

-2.05

呖

-1.51

 khác

-1.37

You

-1.32

esserts

-1.31

 picada

-1.28

In

-1.28

士兵

-1.27

POSITIVE LOGITS

houden

1.45

ãy

1.38

kový

1.38

👼

1.38

 Deze

1.34

歺

1.34

سباب

1.34

CommonModule

1.30



1.30

когда

1.29

Activations Density 0.013%