INDEX

Explanations

its, liver, paper, country, comic, study, dog, guardian, repository

The neuron strongly responds to short, common grammatical function words—especially articles and pronouns (like “a,” “his,” “divided,” etc.).

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

నులు

0.61

 ხელისუფ

0.61

കാരം

0.60

িলারা

0.59

 एजेंस

0.57

IAN

0.57

大きい

0.56

ena

0.56

まで

0.56

까지

0.55

POSITIVE LOGITS

 함수의

1.53

产品的

1.41

项目的

1.41

设备的

1.41

游戏的

1.38

程序的

1.36

文件的

1.34

模型的

1.34

系统的

1.33

数据的

1.33

Activations Density 0.654%