INDEX

Explanations

dividing into categories

The neuron responds to categorical labels or class‐type terms (e.g. “intrinsic,” “nonvolatile,” “primary,” “secondary,” “peripheral,” “central,” “superficial,” “deep,” etc.) indicating classification or subtype.

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

܆

-1.76

出了一

-1.53

issent

-1.52

Même

-1.44

 cappella

-1.43

 équipement

-1.40

izability

-1.39

 veda

-1.39

ctions

-1.38

 همچنین

-1.35

POSITIVE LOGITS

or

2.50

 because

2.31

and

1.91

if

1.87

There

1.57

If

1.53

 или

1.52

 There

1.49

 such

1.47

 этого

1.43

Activations Density 0.451%