INDEX

Explanations

numeric identifiers and values related to databases and data processing

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 7-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_7/width_16k/average_l0_92

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.7.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

 saites

-0.44

 Meksiku

-0.44

 arşivlendi

-0.41

 NDEBUG

-0.40

 estekak

-0.36

oblig

-0.36

 članak

-0.36

 Ее

-0.35

 chủ

-0.35

toHaveBeen

-0.35

POSITIVE LOGITS

 ujednoznacz

0.56

to

0.54

To

0.54

To

0.54

setTo

0.54

DIPSETTING

0.47

 Goto

0.46

archiviato

0.46

titleMargin

0.45

曖昧さ回避

0.45

Activations Density 0.467%

numeric identifiers and values related to databases and data processing

No Known Activations