INDEX

Explanations

proper nouns, specifically names and titles

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

es

-0.48

ians

-0.44

ised

-0.42

ล

-0.42

ました

-0.40

ihin

-0.38

esinde

-0.38

krát

-0.37

iar

-0.37

kh

-0.36

POSITIVE LOGITS

 المعيارى

0.63

transQ

0.58

<?

0.54

\{\\

0.53

unfinished

0.52

 مشين

0.49

ApiModelProperty

0.48

toHaveBeenCalled

0.47

Portail

0.47

parsedMessage

0.47

Activations Density 0.363%

proper nouns, specifically names and titles

No Comments

No Known Activations