INDEX

Explanations

references to specific studies or methodologies in scientific contexts

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

optant type-setters

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

 saites

-0.60

 Италијани

-0.59

 فريبيس

-0.58

<unused80>

-0.58

NameInMap

-0.58

<unused42>

-0.57

<unused52>

-0.57

<unused16>

-0.57

<pad>

-0.57

<unused14>

-0.57

POSITIVE LOGITS

 Photocase

0.34

AutoModerator

0.32

czegó

0.28

 nemlig

0.28

ekš

0.27

 Instyt

0.27

rededor

0.27

 Autorizaciones

0.26

impresa

0.26

accro

0.26

Activations Density 0.008%

references to specific studies or methodologies in scientific contexts

optant type-setters

No Comments

No Known Activations