INDEX

Explanations

list markers

This neuron detects asterisks used as bullet‐point or list‐item markers.

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

 atau

-1.93

玛

-1.76

 became

-1.75

 could

-1.73

 insisted

-1.70

 urged

-1.65

 remained

-1.61

できれば

-1.57

Only

-1.54

">

-1.52

POSITIVE LOGITS

鼍

1.71

“

1.63

 userModel

1.59

 reag

1.58

 Dinas

1.56

};*/

1.54

tA

1.50

出了一

1.49

 voltak

1.48

 звезды

1.45

Activations Density 0.053%