INDEX

Explanations

terms related to blindness and oversight

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

opers

-0.07

eru

-0.07

hell

-0.07

lean

-0.07

ept

-0.07

sto

-0.06

agedList

-0.06

ismatic

-0.06

cla

-0.06

eters

-0.06

POSITIVE LOGITS

ness

0.09

-eye

0.07

nes

0.07

éłĪ

0.07

ening

0.07

ardo

0.07

Sharper

0.06

ãĥ³ãĥĩ

0.06

spe

0.06

fold

0.06

Activations Density 0.007%