INDEX

Explanations

tools and naked eyes

The neuron responds to mentions of concrete tools, implements, or instruments (e.g. marker, shovel, fork, hammer).

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

瑧

-1.09

świę

-1.05

 kabát

-1.04

 лука

-1.00

 stratég

-0.99

источник

-0.97

 sosok

-0.96

＊＊＊＊

-0.96

 hosszú

-0.96

 universitaire

-0.95

POSITIVE LOGITS

to

1.34

就可以

1.19

 Coming

1.10

での

1.01

Coming

0.98

sieu

0.96

unque

0.92

による

0.92

<0xAF>

0.91

Activations Density 0.143%