INDEX

Explanations

relative abundance

The neuron fires on evaluative or emphatic opinion statements (e.g. personal judgments and praise).

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

 bertujuan

0.71

 described

0.70

 embracing

0.68

.\\

0.66

 describing

0.66

:\\

0.66

 utilising

0.63

 analysing

0.63

 באמצעות

0.62

 providing

0.61

POSITIVE LOGITS

 Anna

0.67

0.63

Am

0.62

 Expectations

0.60

0.59

Anna

0.59

 アン

0.57

Ａ

0.57

 Kann

0.56

 Promo

0.56

Activations Density 0.011%