INDEX

Explanations

Math questions and answers

New Auto-Interp

Configuration

Dataset (Dashboard)

Various

Features

131,072

Data Type

float32

Hook Name

blocks.11.hook_resid_post

Architecture

standard

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 copied

-0.07

Bog

-0.07

 Margaret

-0.06

rip

-0.06

DEM

-0.06

Fat

-0.06

čast

-0.06

POSITIVE LOGITS

 osobních

0.07

 heraus

0.06

イ

0.06

 osobní

0.06

defaultValue

0.06

Steps

0.06

andidate

0.06

skb

0.06

 pillars

0.06

 Prime

0.06

Activations Density 0.108%