INDEX

Explanations

elements related to classification methods and hierarchical structures in data processing

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

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

RegressionTest

-0.62

Hentet

-0.57

StoryboardSegue

-0.50

COUVER

-0.48

lujah

-0.48

poche

-0.46

Clik

-0.44

 allergen

-0.44

abank

-0.44

 snippetHide

-0.43

POSITIVE LOGITS

 depending

0.62

Depending

0.47

 Depending

0.47

 whichever

0.43

depending

0.43

 dependiendo

0.40

 ratio

0.40

 kohdetta

0.39

 split

0.38

selectBy

0.38

Activations Density 0.113%

elements related to classification methods and hierarchical structures in data processing

No Comments

No Known Activations