INDEX

Explanations

instead of

The neuron detects contrastive or alternative-introducing expressions—phrases like “instead of,” “rather than,” or similar constructions that set up alternatives.

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

and

-1.54

sandalia

-1.32

 himo

-1.31

屨

-1.29

 also

-1.23

is

-1.22

etheless

-1.22

IBLIO

-1.19

 because

-1.16

作り方

-1.16

POSITIVE LOGITS

we

1.92

 your

1.84

かわらず

1.45

you

1.41

it

1.39

()?;

1.38

เรา

1.36

或

1.36

 chaque

1.34

 allemaal

1.34

Activations Density 0.036%