INDEX

Explanations

citations and references

The neuron is selectively firing on scholarly citation markers—i.e. the bracketed reference numbers (the “[…]” tokens around bibliography or footnote citations).

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

 Some

-2.11

use

-2.06

or

-2.05

It

-1.98

 with

-1.96

This

-1.91

 There

-1.88

 first

-1.85

For

-1.84

If

-1.78

POSITIVE LOGITS

 التطبيق

1.72

woj

1.71

 voisins

1.66

 ſou

1.62

 nouveaux

1.58

 frères

1.58

 quatre

1.55

ọ

1.52

OOOO

1.51

 jeux

1.51

Activations Density 0.010%