INDEX

Explanations

concepts related to privacy and confidentiality

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-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 auroit

-0.58

 useRouter

-0.52

 étoit

-0.50

 enfans

-0.50

 esprits

-0.47

 polaire

-0.47

 tiens

-0.47

 Administrativna

-0.46

 utafitiHapana

-0.46

#+#

-0.46

POSITIVE LOGITS

 privacy

1.02

 Privacy

0.86

 PRIVACY

0.80

privacy

0.77

Privacy

0.77

 secluded

0.70

隐私

0.66

PRIVACY

0.65

 privacidad

0.64

 private

0.59

Activations Density 0.003%

concepts related to privacy and confidentiality

No Comments

No Known Activations