log entry timestamps

The neuron fires on standalone numeric tokens (e.g. version numbers, indices, issue or bibliography numbers).

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

-1.23

 polícia

-1.13

providedIn

-1.09

Preparazione

-1.08

がありますが

-1.06

pleft

-1.06

我对

-1.04

ONLY

-1.02

 escra

-1.02

ributed

-1.02

POSITIVE LOGITS

1.84

1.78

1.73

1.70

1.68

1.59

1.55

1.42

1.33

1.18

Activations Density 0.038%