INDEX

Explanations

dates, places, and entities

The neuron spotlights sequences of short, all-caps or title-cased category or section labels (like “Travel . Food,” “Explore | Connect | Grow,” or “biotech • healthcare • life sciences”) typically found in navigation menus or header bars.

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

 when

-1.92

if

-1.90

 during

-1.74

 before

-1.68

 while

-1.66

in

-1.61

 after

-1.60

 other

-1.50

 where

-1.48

 provide

-1.39

POSITIVE LOGITS

 evaluaciones

1.47

 posiblemente

1.41

 oxígeno

1.34

 sanciones

1.33

 unſ

1.28

 sentimientos

1.24

提出的

1.23

 preguntó

1.21

 conversaciones

1.20

 hapless

1.20

Activations Density 0.045%