INDEX

Explanations

beneath, under, below

The neuron lights up on semantically important “content” words (nouns, verbs, adjectives) rather than on common function words.

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

centerY

-0.79

ギュラー

-0.74

rassed

-0.73

˥

-0.70

arrer

-0.67

 Recap

-0.67

ługo

-0.67

orle

-0.66

outside

-0.65

defendant

-0.65

POSITIVE LOGITS

 beneath

5.63

 underneath

5.56

 under

4.38

 below

4.31

 debajo

4.25

 Beneath

4.16

Beneath

3.94

 زیر

3.73

 dưới

3.50

below

3.27

Activations Density 0.241%