INDEX

Explanations

punctuation following questions or exclamations

The neuron fires strongly on question‐mark punctuation, i.e. it detects “?” tokens.

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

 estou

0.59

gamb

0.58

 Gambling

0.56

ቃል

0.55

 तुम

0.54

 istiyorum

0.54

sales

0.54

Yi

0.53

그리고

0.53

 lasagna

0.53

POSITIVE LOGITS

 wenige

0.45

CTOGRAM

0.44

 درجہ

0.43

他和

0.43

×

0.42

辨

0.42

(=

0.42

 Easily

0.42

 Debye

0.41

他对

0.41

Activations Density 0.066%