INDEX

Explanations

person followed by who

The neuron fires strongly on words naming people or groups of people (e.g. “child,” “children,” “men,” “folks,” “authors”).

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

Where

-1.42

return

-1.38

itu

-1.37

นั้น

-1.36

 bahwa

-1.27

the

-1.21

 quella

-1.20

 serupa

-1.19

 ότι

-1.16

 similar

-1.12

POSITIVE LOGITS

who

11.75

who

4.94

 quien

4.63

 quienes

4.19

 который

4.00

 الذي

4.00

 whom

3.94

 whose

3.75

Who

3.66

 που

3.63

Activations Density 0.437%