INDEX

Explanations

references to locations and geographic names

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 38-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_38/width_16k/average_l0_128

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.38.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

vraagd

-0.39

DC

-0.39

 aspetta

-0.36

 Bedingungen

-0.34

 WASHINGTON

-0.33

 Washington

-0.33

antası

-0.32

 pesta

-0.32

 puesta

-0.31

енча

-0.31

POSITIVE LOGITS

<?

0.57

出版年

0.52

 resourceCulture

0.52

 Quanti

0.51

 quanti

0.50

DockStyle

0.49

Quanti

0.48

okes

0.48

windowFixed

0.48

modelBuilder

0.47

Activations Density 0.354%

references to locations and geographic names

No Comments

No Known Activations