INDEX

Explanations

blocking, filtering, or preventing passage

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

assertIn

-0.77

grupos

-0.77

ادن

-0.77

 ער

-0.77

şen

-0.75

胁

-0.75

 avoidance

-0.73

izacao

-0.73

 verla

-0.73

 stimulating

-0.72

POSITIVE LOGITS

 blocking

2.31

 block

1.95

 filtering

1.85

 blocked

1.80

 filter

1.80

 Blocking

1.76

blocking

1.73

 prevent

1.71

 intercept

1.69

Blocking

1.68

Activations Density 0.136%