INDEX

Explanations

the repetition of the word "same"

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

 similar

-0.07

chos

-0.07

own

-0.07

/fa

-0.06

lan

-0.06

 Similar

-0.06

ÑģÑı

-0.06

 simil

-0.06

Similar

-0.06

POSITIVE LOGITS

-sex

0.15

 thing

0.12

ãģı

0.09

 exact

0.09

 kind

0.09

 zamanda

0.08

-sama

0.08

-origin

0.08

ething

0.07

-old

0.07

Activations Density 0.042%