INDEX

Explanations

could not

This neuron fires on section headings or promotional headlines (short uppercase-styled titles announcing actions or calls to action).

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

鼱

-0.74

assert

-0.72

هي

-0.71

enschappen

-0.70

декс

-0.69

berta

-0.69

 Kampagne

-0.69

ு

-0.68

">(

-0.68

idium

-0.68

POSITIVE LOGITS

 couldn

7.59

 Couldn

5.59

Couldn

5.44

couldn

5.19

 couldnt

4.59

 could

3.17

 konnte

2.56

could

2.31

Could

2.30

 unable

2.25

Activations Density 0.261%