INDEX

Explanations

headline starters

The neuron detects section or category headers—words or phrases followed by a colon, as in headlines or topic labels.

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

 мое

-1.46

-1.38

 unsere

-1.33

鬪

-1.27

什么的

-1.24

 desenli

-1.23

所说的

-1.23

もちろん

-1.23

伤感

-1.23

的时间

-1.20

POSITIVE LOGITS

 from

1.76

it

1.67

how

1.27

ökning

1.25

 where

1.25

affaires

1.25

 売り

1.23

 специфи

1.22

ikiran

1.22

 with

1.22

Activations Density 0.035%