INDEX

Explanations

common punctuation

The neuron strongly activates on sentence-final periods (full-stop punctuation).

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

the

-1.79

is

-1.73

 طف

-1.38

땐

-1.34

 représentants

-1.34

 Geset

-1.33

and

-1.33

 głó

-1.24

 mieszkańców

-1.22

POSITIVE LOGITS

1.51

);

1.47

 étan

1.46

 endommag

1.43

 тот

1.39

或

1.38

 étag

1.38

 éter

1.37

 этих

1.36

].

1.35

Activations Density 0.009%