INDEX

Explanations

VS Code, CSS, CDS Hooks

The neuron strongly activates on uppercase abbreviations and acronyms (e.g., organization codes, flight designations, and other named-entity labels).

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

coprotein

0.64

 oftent

0.64

och

0.61

 видом

0.59

clearly

0.59

cu

0.59

ruik

0.58

वर्ष

0.57

ofthe

0.57

就连

0.57

POSITIVE LOGITS

 gần

0.60

isto

0.57

ющихся

0.54

 derail

0.54

گران

0.53

Ais

0.53

 fortunate

0.53

ที

0.52

ève

0.51

низ

0.51

Activations Density 0.899%