INDEX

Explanations

thank you

The neuron responds strongly to the word “Thank” (capitalized), effectively detecting expressions of thanks or gratitude.

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

 nailing

0.61

 नॉट

0.59

 dvar

0.59

 zählt

0.57

 indel

0.54

 সমু

0.54

 negeri

0.54

 nailed

0.53

स्तव

0.52

 กระ

0.52

POSITIVE LOGITS

tank

0.70

Tank

0.64

 tank

0.61

 Tank

0.59

 cautiously

0.59

type

0.58

公元

0.55

atile

0.55

 favorably

0.55

atar

0.55

Activations Density 0.067%