INDEX

Explanations

contextual headings

The neuron strongly activates on numerical and quantitative information—years, counts, ratings, and review statistics.

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

 without

-1.54

 until

-1.53

in

-1.43

 during

-1.42

can

-1.35

 with

-1.34

 During

-1.30

During

-1.29

 could

-1.25

or

-1.23

POSITIVE LOGITS

 diriger

1.59

 accompagner

1.50

 comptes

1.44

 magnifiques

1.41

〬

1.41

 göm

1.40

髫

1.39

 rassemb

1.38

 faciliter

1.36

 occuper

1.34

Activations Density 0.008%