INDEX

Explanations

"answer"

The neuron activates on instances of the phrase “The answer,” marking where an explicit answer or solution is introduced.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

1.41

1.38

serrat

1.30

yani

1.27

্লাহ

1.25

xanth

1.24

ch

1.23

gallon

1.20

POSITIVE LOGITS

ку

1.22

 وأن

1.20

 answers

1.13

 الاجابه

1.07

ير

1.05

ری

1.05

кою

1.05

 Answers

1.04

 وان

1.02

 tegas

1.02

Activations Density 0.027%