INDEX

Explanations

instances of focusing attention and concentration on a particular topic

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

CodeDom

-0.68

ன்ற

-0.66

 tubing

-0.62

 Martino

-0.62

fde

-0.62

nb

-0.61

ppar

-0.60

umato

-0.60

lblNewLabel

-0.60

 Schalke

-0.60

POSITIVE LOGITS

 focus

2.41

 Focus

2.14

focus

2.09

Focus

2.08

FOCUS

2.06

 FOCUS

2.06

 focuses

2.03

 focused

2.02

 focusing

1.94

 focussed

1.91

Activations Density 0.556%