INDEX

Explanations

unacceptable and excuses

The neuron detects intense negative judgment or condemnation words (e.g. “shocking,” “inexcusable,” “unacceptable,” “criminal”).

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

 jeden

-1.38

 muun

-1.28

who

-1.27

-1.26

 reflexión

-1.25

딕

-1.24

 andern

-1.23



-1.23

Dès

-1.22

Chaque

-1.21

POSITIVE LOGITS

形势

1.38

 שהוא

1.25

and

1.21

𝒔

1.16

־ה

1.14

咐

1.14

 зараз

1.13

 古风

1.13

raisemb

1.12

 that

1.11

Activations Density 0.029%