INDEX

Explanations

verbs of speech or reporting

The neuron fires on words that mark reported speech or speaker attributions (e.g. ask, respond, tell, write, plus the associated names/titles).

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

下一篇

-1.37

-1.34

Mitar

-1.27

đầm

-1.23

如果是

-1.21

 обеспечи

-1.20

Läh

-1.19

 daudz

-1.18

anglicky

-1.17

tattoo

-1.16

POSITIVE LOGITS

/***

1.46

 trent

1.38

Тарих

1.36

糴

1.34

 violon

1.30

(';

1.26

ERIE

1.24

 Cualquier

1.24

 vicin

1.23

멩

1.22

Activations Density 0.016%