INDEX

Explanations

similarity comparisons

This neuron fires on content words that specify quantities or measurements (e.g. terms like “length,” “resources,” “purchased,” “every”)—i.e. quantifiers and other numeric/descriptive terms.

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

bmatrix

-0.91

ório

-0.81

:+:

-0.80

stackrel

-0.79

eração

-0.78

epen

-0.78

 negocio

-0.78

 Lehmann

-0.77

]));

-0.76

reinforced

-0.76

POSITIVE LOGITS

 this

1.41

 these

1.28

 similar

1.21

今回

1.15

今回の

1.14

 where

1.07

 данном

0.89

 exactly

0.88

 similarly

0.86

はこの

0.85

Activations Density 0.110%