INDEX

Explanations

abstract concepts related to uncertainty and variability in models

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

clid

-0.07

upal

-0.07

hack

-0.07

fid

-0.07

gni

-0.07

Hack

-0.07

oyer

-0.06

osphere

-0.06

ourage

-0.06

ÑģÑĤÐ²Ð¾ÑĢ

-0.06

POSITIVE LOGITS

 class

0.07

 contract

0.07

 Aspect

0.07

 aspect

0.07

 sealed

0.06

Aspect

0.06

 partial

0.06

 unsafe

0.06

 trait

0.06

Activations Density 0.004%