INDEX

Explanations

phrases related to crime and criminal activities

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

 îna

-0.50

 puol

-0.48

 Komunikasi

-0.47

 Zelanda

-0.47

 hangat

-0.47

 Hoffnung

-0.46

ewoo

-0.46

 apă

-0.45

joba

-0.45

pelier

-0.44

POSITIVE LOGITS

 crime

1.16

 Crime

1.00

 crimes

1.00

crime

0.96

 CRIME

0.96

 criminal

0.95

Crime

0.94

 Crimes

0.91

 Criminal

0.89

crimes

0.89

Activations Density 0.280%

phrases related to crime and criminal activities

No Comments

No Known Activations