INDEX

Explanations

variations of interaction

The neuron fires on very common, high‐frequency tokens (frequent function and general content words).

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

ucleotide

-0.83

htiö

-0.78

cipes

-0.76

込め

-0.75

alib

-0.75

ั่น

-0.75

ata

-0.74

ADS

-0.73

 amie

-0.73

nal

-0.71

POSITIVE LOGITS

 interaction

7.34

 interactions

6.94

 interact

6.28

interaction

6.09

 Interaction

6.03

 interacting

5.84

 Interactions

5.84

Interaction

5.81

 interacts

5.78

interactions

5.69

Activations Density 0.262%