INDEX

Explanations

punctuation and numeric representations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 36-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_36/width_16k/average_l0_120

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.36.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

 even

-0.50

 ویکی‌پدیای

-0.46

 these

-0.45

 haunt

-0.45

 THAT

-0.45

ValueStyle

-0.43

 ours

-0.43

󠁢

-0.42

lapsing

-0.42

atán

-0.42

POSITIVE LOGITS

 Италијани

0.68

Aholisi

0.60

delwed

0.53

Personendaten

0.52

uxxxx

0.51

 Theſe

0.50

StructEnd

0.47

oa̍t

0.45

AsUp

0.45

 Among

0.45

Activations Density 0.285%

punctuation and numeric representations

No Comments

No Known Activations