INDEX

Explanations

unification and unity concepts

The neuron fires on occurrences of the “unify” lexeme and its morphological variants (unify, unifies, unified, unifying, unification).

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

щита

-2.14

canakan

-2.06

-2.02

 Hingga

-1.96

-1.91

 same

-1.90

 большая

-1.88

 kuchen

-1.84

 Zuhause

-1.84

 gekenn

-1.82

POSITIVE LOGITS

氽

1.89

栊

1.87

 espr

1.82

 stratégie

1.73

’?

1.71

 determina

1.70

coj

1.69

»،

1.66

 subida

1.66

Activations Density 0.027%