INDEX

Explanations

Determining drug classes

This neuron is strongly activated by HTML/XML entity tokens (especially the “'” apostrophe entity) and other numeric‐style tokens (e.g. standalone digits or citation numbers).

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

pagina

0.74

preneur

0.73

イギリス

0.72

 weihn

0.72

istische

0.71

izzazione

0.71

 ressources

0.69

 entrepreneurial

0.68

세계

0.68

 scoperta

0.67

POSITIVE LOGITS

ﻣ

0.76

 Satur

0.75

 Determining

0.72

fie

0.70

ো

0.70

///

0.69

Ꮺ

0.69

 Until

0.69

 contentType

0.69

 Ngày

0.69

Activations Density 0.002%