INDEX

Explanations

terms related to killing or violence

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

 <>",

-0.65

Jeografia

-0.61

orsese

-0.60

 AssemblyCulture

-0.60

LabelTagHelper

-0.59

posedge

-0.59

 informée

-0.59

 Roskov

-0.59

adaptiveStyles

-0.59

 IndexPath

-0.58

POSITIVE LOGITS

 kill

0.54

 Kill

0.49

Kill

0.47

 KILL

0.44

 kills

0.40

destroy

0.40

 Killing

0.36

Killing

0.36

kill

0.35

KILL

0.34

Activations Density 0.202%

terms related to killing or violence

No Comments

No Known Activations