INDEX

Explanations

count or recognition of susceptible targets

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 anywhere

-0.11

IRO

-0.11

 Caval

-0.10

 weapon

-0.10

 dalle

-0.10

aval

-0.10

jal

-0.10

ala

-0.09

 Kron

-0.09

illes

-0.09

POSITIVE LOGITS

 unlucky

0.16

 random

0.14

 unsus

0.13

 innocent

0.13

 whom

0.12

 targets

0.12

 susceptible

0.11

hap

0.11

 vulnerable

0.11

 unfortunate

0.10

Activations Density 0.003%