INDEX

Explanations

mathematical expressions and structures

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

mathematical analysis and notation

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

 chi̍t

-0.39

SL

-0.31

 ممت

-0.30

Bur

-0.30

DU

-0.29

 Бур

-0.29

SP

-0.29

WT

-0.28

GNU

-0.28

aeg

-0.27

POSITIVE LOGITS

 dAtA

0.65

 ſol

0.62

ſelben

0.61

enablog

0.60

ロウィン

0.60

 SYLLABLE

0.59

 ſei

0.57

🜲

0.57

 zwiſchen

0.57

$_(

0.56

Activations Density 0.263%

mathematical expressions and structures

mathematical analysis and notation

No Comments

No Known Activations