INDEX

Explanations

verbs and numbers

The neuron fires on verbs in their conjugated forms (e.g. “explores,” “joined,” “returned,” “offers,” “trying”).

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

 such

-1.51

],

-1.33

 dolayı

-1.27

such

-1.24

捶

-1.23

guro

-1.23

 Ibid

-1.19

NOWLEDG

-1.18

 історі

-1.14

Ibid

-1.13

POSITIVE LOGITS

for

1.55

to

1.28

in

1.27

 excellente

1.11

at

1.06

 feme

1.06

 venons

1.04

 latine

1.03

 hiver

1.02

 flotte

1.02

Activations Density 0.299%