INDEX

Explanations

freeze and frozen concepts

The neuron fires on occurrences of “freeze” (and its inflected forms like froze, freezing, frozen) 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

⢅

-1.86

言った

-1.72

 sofrer

-1.68

 suminist

-1.63

 позволяет

-1.62

 verdaderamente

-1.61

垍

-1.61

geſ

-1.61

鹀

-1.60



-1.58

POSITIVE LOGITS

—

1.84

会

1.84

1.63

for

1.63

if

1.62

冻

1.61

责任编辑

1.60

</strong>

1.54

Тогда

1.54

Freeze

1.52

Activations Density 0.014%