INDEX

Explanations

phrases that express identity and personal state

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

izde

-0.45

 idéia

-0.45

place

-0.44

CascadeType

-0.44

 néglig

-0.43

 daß

-0.43

every

-0.43

 skut

-0.42

places

-0.42

 Fälle

-0.42

POSITIVE LOGITS

MemoryWarning

0.57

addGap

0.55

am

0.54

 EconPapers

0.54

Биография

0.52

Somos

0.52

 happiest

0.51

 gonna

0.51

Estou

0.51

ConstraintMaker

0.50

Activations Density 0.100%

phrases that express identity and personal state

No Comments

No Known Activations