INDEX

Explanations

appeal for help or court

This neuron detects words and phrases related to making an appeal or plea (e.g., “appeal,” “appealing,” “plea”).

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

 ‘‘

-1.72

 ﬁrst

-1.71

îl

-1.71

umpulan

-1.69

媖

-1.64

 ensured

-1.63

 approximating

-1.61

rierten

-1.61

and

-1.60

 échelle

-1.59

POSITIVE LOGITS

屺

1.77

 deswegen

1.73

because

1.67

cherichia

1.67

to

1.66

for

1.63

岽

1.61

1.60

цтва

1.58

凼

1.58

Activations Density 0.019%