INDEX

Explanations

concepts related to responsibility and consequences for past actions

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ilo

-0.08

Ã¦kke

-0.07

ENN

-0.06

iliar

-0.06

iloc

-0.06

edor

-0.06

_expire

-0.06

iox

-0.06

arak

-0.06

POSITIVE LOGITS

 irreversible

0.10

 cannot

0.09

 once

0.08

 genie

0.08

cannot

0.08

 permanent

0.07

åıªèĥ½

0.07

 permanently

0.07

ä¸įèĥ½

0.07

 damage

0.06

Activations Density 0.029%