INDEX

Explanations

license information

The neuron activates strongly on capitalized identifier‐style words (proper names) such as “Razor,” “License,” and “Solution.”

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

viale

-0.85

胥

-0.84

μένος

-0.83

 Wals

-0.82

 daudz

-0.80

Kok

-0.79

 izman

-0.78

yenne

-0.77

の特徴

-0.76

ukų

-0.76

POSITIVE LOGITS

 indisputable

0.89

 входа

0.86

wendigkeit

0.85

 Pira

0.84

нгредієнти

0.81

disposable

0.80

InjectAttribute

0.80

Θ

0.80

zzarella

0.79

㉕

0.79

Activations Density 0.001%