INDEX

Explanations

phrases that address or focus on the listener

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

odon

-0.07

aler

-0.07

ecute

-0.07

mÃŃt

-0.07

oidal

-0.07

edom

-0.07

à¸Īà¸£

-0.07

tainment

-0.06

amar

-0.06

lrt

-0.06

POSITIVE LOGITS

 know

0.17

know

0.13

 knows

0.10

kn

0.09

 Know

0.09

kon

0.09

 known

0.08

-know

0.08

Know

0.08

çŁ¥

0.08

Activations Density 0.018%