INDEX

Explanations

names starting with Charles

This neuron fires on tokens that are part of named entities—especially proper names of people or organizations.

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

jecture

-0.81

jee

-0.76

不该

-0.75

aldo

-0.75

名称

-0.74

罔

-0.73

Though

-0.73

 Though

-0.72

HQ

-0.72

 furiously

-0.71

POSITIVE LOGITS

Dian

0.83

 visi

0.83

 toki

0.81

 pirata

0.81

guilty

0.81

%@

0.80

Gua

0.77

 profili

0.77

 sice

0.75

 Attribution

0.75

Activations Density 0.028%