INDEX

Explanations

terms related to confidentiality and privacy

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

confidential health, intelligence, and relationships

np_acts-logits-general · gemini-2.5-flash-lite

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

AnimationsModule

-0.75

 BorderSide

-0.75

 avoient

-0.71

 auroit

-0.69

 viewDidLoad

-0.66

eamento

-0.66

 feroit

-0.65

 ſont

-0.65

urlpatterns

-0.65

SizeMode

-0.64

POSITIVE LOGITS

 Confidential

0.89

 confidential

0.87

Confidential

0.83

 confidentiality

0.65

identiality

0.62

 Confidentiality

0.61

 secret

0.54

 confided

0.51

 confid

0.50

 onCancelled

0.48

Activations Density 0.003%

terms related to confidentiality and privacy

confidential health, intelligence, and relationships

No Comments

No Known Activations