INDEX

Explanations

Python code structure and keywords

The neuron responds to calls‐to‐action directing the reader to visit or click a web link (e.g. “go to web site,” “click the following webpage”).

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

тись

0.59

calories

0.54

тири

0.52

বুব

0.52

restaurants

0.52

вшись

0.52

ailure

0.51

casino

0.51

izacin

0.51

 বিপর

0.50

POSITIVE LOGITS

Ā

0.51

 Peirce

0.50

 adduced

0.49

Ж

0.48

Ș

0.47

Ё

0.47

:\

0.46

 Förder

0.46

 McInt

0.46

 Materialien

0.45

Activations Density 0.003%