INDEX

Explanations

elements related to the structure and content of a collection or set of media, particularly highlighting supplementary materials and features

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

AndEndTag

-0.37

 Zoning

-0.37

 social

-0.36

 ModelExpression

-0.35

]-->

-0.35

receita

-0.35

zn

-0.34

 schie

-0.34

jectory

-0.34

 Polic

-0.34

POSITIVE LOGITS

 reissue

0.62

Tembelea

0.52

 հղումներ

0.47

 noDo

0.45

 Biôgrafia

0.44

verifyException

0.43

 remastered

0.43

 Krakowie

0.43

 ofere

0.43

 deluxe

0.43

Activations Density 0.020%

elements related to the structure and content of a collection or set of media, particularly highlighting supplementary materials and features

No Comments

No Known Activations