INDEX

Explanations

questions and relationships

This neuron detects terms associated with educational practice and assessment materials (e.g. worksheets, exams, tests, answer memoranda, cram).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 their

-1.28

its

-1.13

網址

-1.13

 Consejo

-1.13

 oryginal

-1.10

árny

-1.07

it

-1.06

clays

-1.05

 quite

-1.01

ってしまった

-1.01

POSITIVE LOGITS

・

1.20

 eccell

1.11

 tyrants

1.11

erschiedene

1.09

 pourra

1.09

 của

1.09

 зрения

1.08

rouvez

1.07

κτηρισ

1.07

 району

1.06

Activations Density 0.000%