INDEX

Explanations

action verbs

The neuron activates on verbs that denote concrete actions or procedures (e.g. “access,” “configure,” “operate,” “apply,” “perform,” “attend,” “participate,” etc.).

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

了一些

-1.42

 also

-1.41

というのも

-1.38

一点点

-1.34

 autre

-1.33

Поэтому

-1.31

by

-1.30

mientos

-1.30

 інші

-1.30

not

-1.28

POSITIVE LOGITS

 such

2.23

 this

2.14

1.65

 these

1.62

 黄色

1.52

 spør

1.51

such

1.49

its

1.46

 安心

1.45

entrega

1.42

Activations Density 0.141%