INDEX

Explanations

objects in a scene

The neuron fires on concrete, manipulable objects and props (e.g. cups, chairs, blankets, glasses, tools) mentioned in the text.

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

-1.83

 corações

-1.62

 ketika

-1.55

llac

-1.55

orologio

-1.42

 食事

-1.42

 wilden

-1.38

 amikor

-1.36

 these

-1.34

før

-1.34

POSITIVE LOGITS

1.42

1.25

 already

1.22

 więc

1.21

↵↵↵

1.20

ThemeOverlay

1.19

altor

1.16

 redan

1.16

</u>

1.14

Activations Density 0.015%