INDEX

Explanations

references to violent actions or 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 @ 27-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_27/width_16k/average_l0_118

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.27.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

 obreros

-0.40

>{@

-0.38

 huesos

-0.34

(!__

-0.34

 hombros

-0.33

 trock

-0.33

 pogod

-0.32

 Werbung

-0.32

ěr

-0.31

 Schon

-0.31

POSITIVE LOGITS

 guns

1.36

 firearms

1.31

gun

1.31

 firearm

1.27

guns

1.25

 pistol

1.15

 weapons

1.14

 Guns

1.13

Guns

1.13

 Firearms

1.10

Activations Density 0.385%

references to violent actions or criminal activities

No Comments

No Known Activations