INDEX

Explanations

references to specific technical or scientific data sources

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.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

:✨

-0.48

 autorytatywna

-0.44

 peccato

-0.44

 löytyy

-0.43

TLS

-0.42

ię

-0.38

 Insol

-0.38

 tunn

-0.38

 meille

-0.37

miyorum

-0.37

POSITIVE LOGITS

AccessorTable

0.49

 Infór

0.46

jface

0.45

endcsname

0.43

risa

0.41

 actionPerformed

0.41

FailureListener

0.40

zzleHttp

0.40

XDECREF

0.40

 nominal

0.39

Activations Density 0.417%

references to specific technical or scientific data sources

No Comments

No Known Activations