INDEX

Explanations

explaining the cause

The neuron fires strongly on occurrences of the phrase “thanks to,” i.e. expressions of grateful attribution.

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

 colonie

-1.80

🕗

-1.74

⬮

-1.70

 може

-1.70

čnosti

-1.54

 poffe

-1.51

芻

-1.51

 casse

-1.49

韪

-1.48

 occaf

-1.48

POSITIVE LOGITS

our

1.87

his

1.77

 some

1.71

 those

1.66

1.64

Of

1.62

Of

1.59

 przede

1.59

 また

1.55

 with

1.54

Activations Density 0.011%