INDEX

Explanations

references to reports or studies

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

wpi

-0.39

 hadn

-0.35

ceğine

-0.35

werken

-0.34

 lucru

-0.34

 wesentlichen

-0.33

IonicModule

-0.32

 prezent

-0.31

 lessons

-0.31

 jongen

-0.31

POSITIVE LOGITS

httphttps

0.92

 resourceCulture

0.68

 noDo

0.66

 autorytatywna

0.64

 AssemblyTitle

0.62

/*

0.60

BufferException

0.60

 surla

0.59

oredCriteria

0.59

رشف

0.57

Activations Density 0.079%

references to reports or studies

No Comments

No Known Activations