INDEX

Explanations

defining dimensions or coordinates

The neuron never activates—i.e. it does not detect any specific token pattern.

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

or

-1.75

-1.70

if

-1.69

-1.59

 enhances

-1.41

-1.40

-1.33

 which

-1.32

when

-1.30

 reinforces

-1.30

POSITIVE LOGITS

杁

1.80

扨

1.78

ERIA

1.66

 Unterschiede

1.59

拊

1.57

 fazenda

1.52

啐

1.51

嵓

1.51

 sooo

1.49

 agricole

1.49

Activations Density 0.002%