INDEX

Explanations

numerical dimensions and sizes within the text

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

dimensions and measurements

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

 ***!

-0.65

 للاسماء

-0.61

ंदीखरीदारी

-0.61

 pinulongan

-0.57

BufferException

-0.57

 BoxDecoration

-0.56

########.

-0.56

 يتيمه

-0.54

 geſch

-0.54

 المعيارى

-0.54

POSITIVE LOGITS

 resourceCulture

0.36

全長

0.33

ature

0.33

él

0.30

uç

0.29

Durata

0.28

程

0.28

verses

0.27

 uitge

0.27

 apresent

0.27

Activations Density 0.005%

numerical dimensions and sizes within the text

dimensions and measurements

No Comments

No Known Activations