INDEX

Explanations

disqualified or qualified

The neuron activates on various forms and abbreviations of “disqualify” (e.g. disqualified, disqualify, DQ).

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

 Когда

-2.27

</h3>

-2.22

 When

-2.14

 เมื่อ

-2.09

-1.98

 Therefore

-1.91

嵛

-1.91

鞮

-1.90

 жизни

-1.89

#############

-1.89

POSITIVE LOGITS

isation

2.05

主要是

1.92

’

1.88

尨

1.82

zelfde

1.78

</b>

1.77

 файлов

1.73

 使っ

1.71

 красиво

1.67

焄

1.64

Activations Density 0.003%