INDEX

Explanations

ok, fine, alright, better

This neuron detects brief positive evaluation words (e.g. “ok,” “fine,” “alright,” “better”).

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

</b>

-1.66

riezmann

-1.64

 aplicable

-1.59

Ꮈ

-1.56

Ꮵ

-1.55

 delicioso

-1.55

necessary

-1.53

ﺢ

-1.52

 recomendable

-1.48

 musíte

-1.48

POSITIVE LOGITS

so

1.54

1.46

at

1.40

 relacionado

1.34

 from

1.30

’

1.28

1.27

 unlike

1.27

水晶

1.23

The

1.22

Activations Density 0.003%