INDEX

Explanations

python code

Explanation of neuron 4 behavior: the main thing this neuron does is find formal procedural or instructional language (e.g. requirements, permissions, registrations, and “how‐to” steps).

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

0.55

மிட

0.54

']))

0.53

"]))

0.53

॰ऍ

0.52

रेटर

0.52

 ventric

0.51

 Nope

0.51

ంద్

0.50

__."

0.50

POSITIVE LOGITS

キュ

0.59

៩

0.54

 Listed

0.54

 இவற்றை

0.54

 ناول

0.53

இந்தப்

0.51

سا

0.50

лення

0.50

＊

0.50

listed

0.49

Activations Density 0.033%