INDEX

Explanations

future time references

The neuron activates on words and phrases that refer to future time or scheduling (e.g. “later,” “next week,” “Monday”).

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

 rése

-1.42

ORMAL

-1.30

 stratég

-1.30

 solamente

-1.27

 prenn

-1.27

肮

-1.26

 alho

-1.26

AILY

-1.23

 alpes

-1.22

 baut

-1.21

POSITIVE LOGITS

if

1.66

 when

1.52

 then

1.45

 with

1.43

When

1.41

 their

1.41

 which

1.39

 because

1.38

or

1.34

Since

1.34

Activations Density 0.026%