INDEX

Explanations

section headers and organizational cues in structured documents

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

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 everywhere

-0.44

everywhere

-0.39

Климат

-0.39

IntoConstraints

-0.38

Geografie

-0.38

ConstraintMaker

-0.37

 pena

-0.36

astify

-0.36

 الرياضيه

-0.35

you

-0.35

POSITIVE LOGITS

0.59

 tartalomajánló

0.58

Portale

0.56

ंदीखरीदारी

0.54

ReusableCell

0.53

 ExecuteAsync

0.51

pexpr

0.51

 ویکی‌پدیای

0.51

 ویکی‌پدی

0.50

endpush

0.50

Activations Density 0.029%

section headers and organizational cues in structured documents

No Comments

No Known Activations