INDEX

Explanations

"we" followed by verbs

The neuron is looking for occurrences of the first‐person plural pronoun “we.”

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

we

-2.09

but

-1.71

けません

-1.64

.”

-1.63

 même

-1.60

mpty

-1.59

;”

-1.59

,”

-1.59

 ส์

-1.58

was

-1.53

POSITIVE LOGITS

1.68

 позволяют

1.55

 allow

1.54

section

1.50

which

1.43

they

1.42

1.41

 zvíř

1.37

 create

1.37

</em>

1.37

Activations Density 0.012%