INDEX

Explanations

start of code conditions

The neuron fires on named entities (proper nouns, titles) and numerical tokens.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

Whilst

1.91

 Whilst

1.82

 whilst

1.50

 गाँधी

1.46

١

1.40

 Amongst

1.31

 हूँ

1.30

١

1.30

２

1.26

๑

1.24

POSITIVE LOGITS

—

1.10

 toward

1.01

 अा

0.96

 nonprofit

0.91

 yüzde

0.90

ോ

0.88

 nong

0.87

 spokeswoman

0.85

 five

0.84

 지난해

0.84

Activations Density 0.707%