INDEX

Explanations

methods and techniques

The neuron activates on terms that refer to research methods or techniques (e.g. “methods,” “techniques,” “approaches,” “procedures”).

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

-1.55

 capucha

-1.55

コマ

-1.31

đầm

-1.30

ähn

-1.30

窰

-1.30

 their

-1.27

 expérience

-1.27

flä

-1.26

щении

-1.25

POSITIVE LOGITS

to

2.30

 такі

1.61

↵↵

1.49

</em>

1.47

 меры

1.46

 wist

1.45

 Staying

1.42

香り

1.41

 folgende

1.41

 quella

1.41

Activations Density 0.115%