INDEX

Explanations

biography subjects

The neuron primarily activates on numeric tokens and measurements (e.g. years, distances, times, star ratings)—in other words, it detects numbers.

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

არი

-0.71

Provider

-0.69

 연결

-0.69

 provider

-0.68

ORB

-0.68

Mes

-0.68

 belief

-0.68

 homemaker

-0.67

 homestead

-0.66

bibinfo

-0.66

POSITIVE LOGITS

対戦

0.71

BUF

0.71

뽐

0.68

 Ranks

0.67

 Galloway

0.66

 Oviedo

0.65

 سر

0.64

Tok

0.64

 CONCER

0.63

duma

0.62

Activations Density 0.058%