INDEX

Explanations

clearance and interference

The neuron is activated by numeric tokens, especially measurement values and other numbers.

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

wander

-0.91

 wata

-0.82

 failings

-0.82

㎎

-0.81

 beaux

-0.80

 pesante

-0.78

abonnement

-0.77

 hierdie

-0.77

Aplica

-0.77

身上的

-0.76

POSITIVE LOGITS

 clearance

3.38

 Clearance

2.52

clearance

2.44

 clearances

2.19

 interference

1.84

 interferes

1.59

 interfere

1.59

interference

1.58

 clearing

1.54

 Interference

1.49

Activations Density 0.036%