INDEX

Explanations

importance role

The neuron fires on mentions of something “playing a … role” (e.g. “central role,” “key role,” “major role”) indicating the function or importance of an entity.

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

 hugely

-0.96

ló

-0.93

瘟

-0.88

ꪮ

-0.88

 équipé

-0.88

罢

-0.87

 Nutzen

-0.86

 موسیقی

-0.86

他就

-0.85

 jaka

-0.84

POSITIVE LOGITS

 when

1.25

when

1.11

所以在

0.95

 cuando

0.91

خطيط

0.91

themselves

0.89

 during

0.88

in

0.88

 sebagainya

0.88

 сумму

0.88

Activations Density 0.063%