INDEX

Explanations

experimental improvisation

The neuron activates on numeric tokens, i.e. digits and numbers (years, counts, catalog numbers, 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

ksiyon

-0.82

⅛

-0.81

さで

-0.81

 hipó

-0.79

uló

-0.79

 team

-0.77

kies

-0.77

hook

-0.77

 Flach

-0.77

 district

-0.76

POSITIVE LOGITS

 improvisation

1.41

 noise

1.24

 improvis

1.21

Sonic

1.18

 improvised

1.16

 EXPERIMENTAL

1.10

 minimalism

1.09

 experimental

1.08

 Impro

1.07

noise

1.06

Activations Density 0.073%