INDEX

Explanations

the word "off" and its variations in the context of offenses

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 0-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_0/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.0.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

 HasFactory

-0.58

alade

-0.55

BagConstraints

-0.53

 migraines

-0.52

 scrubs

-0.51

complexType

-0.50

empêcher

-0.50

 estrenar

-0.49

 paddles

-0.49

 devaluation

-0.49

POSITIVE LOGITS

Off

2.59

Off

2.58

off

2.44

OFF

2.27

off

2.27

OFF

2.22

オフ

1.80

offs

1.70

 offs

1.46

 オフ

1.45

Activations Density 0.211%

the word "off" and its variations in the context of offenses

No Comments

No Known Activations