INDEX

Explanations

take advantage / precautions / liberties

The neuron strongly activates on occurrences of the verb “take” (including “takes” and “taking”).

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

א

1.45

ع

1.29

 aptly

1.26

س

1.20

вата

1.20

ಾರ್ಥ

1.16

зда

1.10

 echoed

1.10

폰

1.10

 squeezed

1.08

POSITIVE LOGITS

 Threats

1.26

 নেয়

1.24

ación

1.22

 Liberties

1.22

 নিচ্ছে

1.20

 ambil

1.20

ro

1.20

 Prend

1.20

 Took

1.18

 Turns

1.17

Activations Density 0.167%