INDEX

Explanations

movement and past actions

The neuron strongly activates on verbs of motion, especially forms of “go,” “went,” and “come.”

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

erráneo

-1.81

ývá

-1.64

MeasureSpec

-1.52

izielle

-1.43

 GenerationType

-1.41

 которы

-1.41

ኋ

-1.40

 votre

-1.39

 mikrofon

-1.38

␉

-1.38

POSITIVE LOGITS

for

2.02

was

1.75

of

1.70

to

1.58

1.56

 only

1.53

 already

1.50

his

1.42

 said

1.42

Activations Density 0.016%