INDEX

Explanations

keywords related to secrets and forbidden knowledge

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.18.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

beiter

-0.52

 Gila

-0.52

 Urugu

-0.51

 Arjuna

-0.51

 Türkçe

-0.51

 UITableViewCell

-0.51

XmlAccessorType

-0.50

 Cambo

-0.50

parsedMessage

-0.50

 Hamas

-0.50

POSITIVE LOGITS

 fictional

0.63

 town

0.57

 Callahan

0.56

 neighbouring

0.51

 Paglinawan

0.50

 متحده

0.49

Nowhere

0.49

NewReader

0.48

 royaume

0.48

 City

0.48

Activations Density 0.259%

keywords related to secrets and forbidden knowledge

No Comments

No Known Activations