INDEX

Explanations

still,

This neuron activates on the discourse connective “still,” marking a contrasting or continuing thought.

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

↵↵

-2.02

 From

-1.70

and

-1.68

はい

-1.62

зно

-1.59

is

-1.58

At

-1.54

nemer

-1.54

谖

-1.51

atı

-1.49

POSITIVE LOGITS

 Vikipedi

1.79

работа

1.75

iertos

1.75

 เป็น

1.73

 selben

1.70

 möge

1.70

 wouldn

1.68

]")]

1.63

 ersche

1.62

 desselben

1.60

Activations Density 0.007%