INDEX

Explanations

references to different editions of works or publications

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

Not in Any Lists

No Comments

Negative Logits

 estekak

-0.60

 beginnetje

-0.58

ThroughAttribute

-0.57

 pinulongan

-0.57

✨:

-0.52

 sexuelle

-0.50

OGND

-0.48

 publicitaires

-0.47

//

-0.46

地说道

-0.46

POSITIVE LOGITS

 edition

2.59

 Edition

2.50

Edition

2.22

 EDITION

2.19

 editions

2.14

edition

2.13

 Editions

1.63

 édition

1.58

 edición

1.48

 edição

1.39

Activations Density 0.134%

references to different editions of works or publications

No Comments

No Known Activations