INDEX

Explanations

assessing things as good or bad

The neuron is tuned to detect occurrences of the phrase “good thing” (often as “a good thing”).

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

']='

-1.46

 from

-1.39

 máme

-1.25

 presentará

-1.24

нью

-1.24

мови

-1.24

፨

-1.23

陨

-1.22

anguardia

-1.21

POSITIVE LOGITS

↵↵

2.02

1.63

别人

1.47

1.45

1.39

 reason

1.38

<b>

1.35

1.31

1.30

fully

1.30

Activations Density 0.032%