INDEX

Explanations

responsibility for accuracy

The neuron detects disclaimer language around “responsible,” especially in “not responsible for” or “responsible for” legal-sounding statements.

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

for

-1.55

 from

-1.41

libat

-1.34

to

-1.33

'...

-1.30

 like

-1.27

 reduced

-1.27

There

-1.24

During

-1.24

After

-1.21

POSITIVE LOGITS

ception

1.41

fusions

1.40

 muebles

1.40

曁

1.35

ícias

1.33

viembre

1.31

</u>

1.30

layah

1.29

ges

1.26

辔

1.26

Activations Density 0.048%