INDEX

Explanations

various items in lists

The neuron activates on named-entity tokens—proper names, labels, or tags (e.g., country names, band or show titles, HTML element names).

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

 that

-1.20

 wied

-1.16

おわりに

-1.13

—

-1.13

⇌

-1.12



-1.12

_();

-1.09

出来た

-1.07

POSITIVE LOGITS

各种

1.41

各種

1.30

许多

1.11

 många

1.11

 første

1.09

 menangis

1.07

 arrivée

1.07

<bos>

1.05

两种

1.05

課程

1.05

Activations Density 0.044%