INDEX

Explanations

specific nouns and context modifiers

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 bashing

1.38

 زمینه

1.23

ərb

1.22

Calculation

1.22

Percentages

1.22

咗

1.21

 शरण

1.20

δες

1.18

 janë

1.18

 سسٹم

1.17

POSITIVE LOGITS

 from

1.28

 between

1.27

 whose

1.17

 within

1.15

 with

1.10

 that

1.04

 which

0.98

 this

0.98

for

0.98

 across

0.97

Activations Density 0.287%