INDEX

Explanations

percentage-related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 13-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_13/width_65k/average_l0_75

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.13.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

 Horner

-0.91

 Paglinawan

-0.84

 resourceCulture

-0.81

JsonFormat

-0.81

 estimés

-0.79

FormTagHelper

-0.79

 HORN

-0.76

 Horowitz

-0.76

Hor

-0.75

ViewFeatures

-0.75

POSITIVE LOGITS

%%

1.00

%%

0.99

Gim

0.94

Gim

0.81

Gimme

0.77

ValueStyle

0.76

 Zwie

0.75

 Sevilla

0.74

 Pilipinas

0.73

omsday

0.72

Activations Density 0.038%

percentage-related terms

No Comments

No Known Activations