INDEX

Explanations

numbers and symbols

The neuron consistently activates on standalone numeric tokens (e.g. “72.5,” “60,” “65.5”), indicating it’s looking for numbers or numerical measurements in the text.

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

 with

-2.41

 from

-2.16

 that

-1.97

 poignant

-1.84

by

-1.75

 frantic

-1.71

for

-1.70

 While

-1.69

 when

-1.64

 will

-1.59

POSITIVE LOGITS

ఱ

1.80

 praticamente

1.74

 tivi

1.72

 soooo

1.67

ほとん

1.66

ort

1.61

 khong

1.60

oe

1.59

 amazing

1.59

 sooo

1.58

Activations Density 0.131%