INDEX

Explanations

pronouns and associated feelings/actions

The neuron activates on words describing how a text affects its audience—e.g. tokens expressing what “you” or “people” “make,” “feel,” or “think.”

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

HEAP

0.57

 Judgment

0.55

 Gallup

0.55

 tỏ

0.54

უფ

0.53

 JOURNAL

0.53

ibir

0.52

 kaip

0.52

 Arbitration

0.52

 Acceptance

0.51

POSITIVE LOGITS

 rethink

0.94

 rethinking

0.91

 reconsider

0.83

 सोचने

0.76

 задума

0.72

 chuckle

0.70

 chuckled

0.70

 wonder

0.66

 উজ্জ

0.65

 speechless

0.64

Activations Density 0.297%