INDEX

Explanations

stimulus/input/key/factor relationship

This neuron fires strongly on multi‐syllable, content-bearing nouns (especially large-scale or abstract nouns).

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

 เพราะ

0.84

 porque

0.76

 क्योंकि

0.75

!!!!

0.73

 کیونکہ

0.72

!!

0.69

!!!

0.68

 because

0.66

 perché

0.66

 因為

0.66

POSITIVE LOGITS

pecific

0.80

such

0.71

 such

0.70

そのような

0.68

 Athenian

0.67

تان

0.65

tä

0.64

 SUCH

0.64

そういう

0.63

 mentored

0.63

Activations Density 0.001%