INDEX

Explanations

references to attacks and violence

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 22-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_22/width_16k/average_l0_72

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.22.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

СТВА

-0.63

GEBURTSDATUM

-0.58

trl

-0.56

("$.

-0.55

OOO

-0.54

caine

-0.54

haltens

-0.54

guli

-0.54

Bett

-0.53

NESS

-0.53

POSITIVE LOGITS

 Attacks

0.78

Attacks

0.72

 attacks

0.69

attacks

0.68

 ATTACK

0.66

ATTACK

0.64

 MessageBoxIcon

0.62

 Attack

0.61

WireFormatLite

0.60

 Taktlose

0.60

Activations Density 0.120%

references to attacks and violence

No Comments

No Known Activations