INDEX

Explanations

hedging words and conditional phrases

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Func

1.11

 Funktions

1.05

функциона

1.02

 dysfunction

1.01

 Functional

0.97

 Horn

0.95

 Norm

0.94

functioning

0.91

Sic

0.90

 Function

0.90

POSITIVE LOGITS

 outweigh

1.33

 outweighs

1.17

 intang

1.17

 instilled

1.15

 rekind

1.14

浓

1.12

 manifested

1.11

 embodied

1.10

 outweighed

1.08

 intangible

1.07

Activations Density 0.523%