INDEX

Explanations

compound nouns and descriptive phrases

The neuron is triggered by major topical or domain‐specific keywords and proper‐noun phrases—e.g. the names of papers, books, technical terms, or article titles.

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



-1.48

一

-1.47

 things

-1.44

 Perſ

-1.41

 unbelievably

-1.38

 manguera

-1.34

 estranho

-1.34

 znajdu

-1.32

 ſtre

-1.30

 bén

-1.30

POSITIVE LOGITS

and

1.92

on

1.46

at

1.43

in

1.39

 with

1.28

相关文章

1.25

_{-}

1.24

（

1.22

”，

1.21

诸多

1.21

Activations Density 0.596%