INDEX

Explanations

you'll see

The neuron is picking up instances of the verb “see” (as in “you’ll see,” “see the following,” etc.) in instructional or descriptive contexts.

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

if

-2.33

and

-1.97

 wants

-1.84

it

-1.70

 sleek

-1.69

 what

-1.67

or

-1.64

 make

-1.63

 give

-1.55

 provides

-1.55

POSITIVE LOGITS

1.66

などが

1.52

contentValues

1.44

1.34

Might

1.33

 erinner

1.32

 traductions

1.28

 erwar

1.27

 lastly

1.27

Activations Density 0.029%