INDEX

Explanations

asking questions and inquiring

This neuron fires on instances of the verb “ask” (e.g. “ask,” “asked”) indicating someone making a request or posing a question.

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

-2.45

-2.31

-1.96

-1.77

-1.75

Which

-1.73

 noen

-1.64

</b>

-1.60

</h2>

-1.59

POSITIVE LOGITS

းက

1.73

籐

1.63

Donativos

1.60

</sub>

1.59

 from

1.55

 EXERC

1.50

?」

1.49

拳头

1.48

’?

1.48

 mildly

1.48

Activations Density 0.004%