INDEX

Explanations

lists, delimiters, headings

The neuron fires on colons that introduce document metadata fields or section headings (e.g. “Aim:”, “Skill level:”, “Keywords:”, “Dosage and Administration:”).

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

</strong>

-2.73

-2.33

</u>

-2.02

</code>

-1.91

</h4>

-1.84

be

-1.82

他的

-1.82

ING

-1.81

嗵

-1.79

POSITIVE LOGITS

you

2.41

but

2.17

be

1.99

 Anſ

1.93

了你

1.88

訁

1.86



1.86

 essas

1.86

 mencionados

1.77

Ᏹ

1.74

Activations Density 0.002%