INDEX

Explanations

proper nouns related to locations and institutions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 34-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_34/width_16k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.34.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

UrlResolution

-0.68

OGND

-0.60

 виправивши

-0.59

 HFILL

-0.57

WebVitals

-0.56

 المعيارى

-0.56

 ویکی‌پدیا

-0.54

ädie

-0.52

DrawerToggle

-0.52

XmlAccessType

-0.52

POSITIVE LOGITS

 Notre

0.52

Notre

0.46

ND

0.46

khart

0.40

 Niles

0.39

ND

0.38

Mic

0.38

 Dame

0.36

Pok

0.35

hiana

0.35

Activations Density 0.202%

proper nouns related to locations and institutions

No Comments

No Known Activations