INDEX

Explanations

Gene expression analysis

New Auto-Interp

Configuration

Dataset (Dashboard)

Various

Features

131,072

Data Type

float32

Hook Name

blocks.27.hook_resid_post

Architecture

standard

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 foster

-0.07

achts

-0.07

姓

-0.07

 saat

-0.07

٬

-0.06

＊

-0.06

blem

-0.06

 vente

-0.06

 satu

-0.06

rescia

-0.06

POSITIVE LOGITS

sdk

0.06

_DAY

0.06

_PART

0.06

 entrar

0.06

Any

0.06

.fromFunction

0.06

 DARK

0.06

Considering

0.06

contest

0.06

-Agent

0.06

Activations Density 0.015%

No Known Activations