INDEX

Explanations

structured numerical data or sequences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.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

 ainfi

-0.75

 plufieurs

-0.74

 miniaturka

-0.70

 négo

-0.68

 feroit

-0.66

 heureuse

-0.66

 boneca

-0.64

 caneca

-0.63

 anún

-0.63

 sprawy

-0.63

POSITIVE LOGITS

 Dons

0.63

ocs

0.59

ODS

0.59

IOD

0.58

})}

0.56

OGND

0.56

poch

0.55

')"

0.55

<>

0.55

poc

0.54

Activations Density 0.811%

structured numerical data or sequences

No Comments

No Known Activations