INDEX

Explanations

elements related to scientific or academic data analysis

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

ItemBackground

-0.69

EDEFAULT

-0.65

AndEndTag

-0.65

styleable

-0.65

Derbyniad

-0.64

 AttributeSet

-0.62

 '\\;'

-0.62

 InputDecoration

-0.61

 autorytatywna

-0.60

ьаж

-0.60

POSITIVE LOGITS

 begründet

0.41

 Wirklichkeit

0.36

ineſs

0.33

 porcelana

0.32

mę

0.32

 Castilla

0.32

Single

0.31

])]

0.31

ech

0.31

True

0.31

Activations Density 3.953%

elements related to scientific or academic data analysis

No Comments

No Known Activations