INDEX

Explanations

thought

The neuron responds to occurrences of the word “thought.”

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

狃

-1.69

盉

-1.63



-1.59

栒

-1.57

鮰

-1.52

眛

-1.48

いだ

-1.46

Stalin

-1.44

 máme

-1.44

okkal

-1.43

POSITIVE LOGITS

1.78

re

1.70

别说

1.70

Not

1.58

 obtain

1.53

 meticulous

1.51

 boast

1.45

 purchase

1.45

ㆆ

1.44

 even

1.41

Activations Density 0.007%